Models

Muse Wonder Full from Meta AI, language,vision,video,multimodal model published 2026-10-12. Notable for: discretionary.

Muse Wonder from Meta AI, language,vision,video,multimodal model published 2026-10-12. Notable for: discretionary.

Gemini 4 Argon from Google DeepMind, language,multimodal,vision model published 2026-09-30. Notable for: discretionary.

GPT-6.1 Sol from OpenAI, multimodal,language,vision model published 2026-09-29. Notable for: discretionary.

Claude Sonnet 5.5 from Anthropic, language,multimodal,vision model published 2026-09-28. Notable for: discretionary.

GPT-6 Sol from OpenAI, multimodal,language,vision model published 2026-09-22. Notable for: discretionary.

GPT-6 Luna Decisions from OpenAI, language,vision,multimodal model published 2026-09-22. Notable for: discretionary.

GPT-6 Luna from OpenAI, multimodal,language,vision model published 2026-09-22. Notable for: discretionary.

Claude Opus 5.5 from Anthropic, language model published 2026-09-22. Notable for: discretionary.

Grok 4.7 from xAI, language model published 2026-09-21. Notable for: discretionary.

Step 5 Preview from StepFun, language,vision,multimodal model published 2026-09-18. Notable for: discretionary.

DeepSeek V4.1 Flash from DeepSeek, language,vision,multimodal model published 2026-09-09. Notable for: discretionary.

Mercury 2.5 from Inception Labs, language model published 2026-09-08. Notable for: discretionary.

GPT-Image-2.5 Sunburst from OpenAI, image generation model published 2026-09-08. Notable for: discretionary.

GPT-Image-2.5 Flare from OpenAI, image generation model published 2026-09-08. Notable for: discretionary.

ChatGPT Images 2.5 from OpenAI, image generation,vision model published 2026-09-08. Notable for: discretionary.

WeatherNext 3 from Google DeepMind, Google Research, earth science model published 2026-09-03. Notable for: significant use,discretionary,sota improvement.

OpenEvidence Darwin from OpenEvidence, language,medicine model published 2026-09-03. Notable for: sota improvement.

GPT-6 Astra from OpenAI, multimodal,language,vision model published 2026-09-03. Notable for: sota improvement,discretionary,significant use,training cost.

Muse Spark 1.3 from Meta AI, language model published 2026-09-02. Notable for: discretionary,significant use.

Gemini 3.8 Flash from Google DeepMind, language,multimodal,image generation,video,audio model published 2026-09-02. Notable for: discretionary.

Claude Fable 5.1 from Anthropic, language,multimodal,vision model published 2026-09-01. Notable for: discretionary,significant use,sota improvement.

Tencent Hy4 preview from Tencent, language model published 2026-08-28. Notable for: discretionary.

GLM-5.3-Flash from Z.ai (Zhipu AI), language,vision,multimodal model published 2026-08-20. Notable for: discretionary.

GEN-1.5 from Generalist, robotics,language,multimodal model published 2026-08-19. Notable for: discretionary.

GLM-5.3 from Z.ai (Zhipu AI), language model published 2026-08-14. Notable for: discretionary.

GPT-5.6 Cyber from OpenAI, cybersecurity,language model published 2026-08-11. Notable for: discretionary.

Qwen3.8-2.4T-A95B from Alibaba, language model published 2026-08-08. Notable for: discretionary.

Motif-3 from Motif Technologies, language model published 2026-08-07. Notable for: training cost.

Muse Spark 1.2 from Meta AI, language model published 2026-08-05. Notable for: discretionary.

K-EXAONE 2.0 from LG AI Research, language model published 2026-07-31. Notable for: training cost.

A.X K2 from SK Telecom, language model published 2026-07-29. Notable for: training cost.

Claude Opus 5 from Anthropic, language,vision,multimodal model published 2026-07-24. Notable for: sota improvement.

Qwen 3.8 Max from Alibaba, language,vision,video model published 2026-07-19. Notable for: discretionary.

Kimi K3 from Moonshot, language,multimodal,vision model published 2026-07-16. Notable for: discretionary,significant use.

Inkling from Thinking Machines, language,multimodal,audio,video,vision model published 2026-07-15. Notable for: discretionary.

Tencent Hy3 from Tencent, language model published 2026-07-06. Notable for: discretionary.

LongCat-2.0 from Meituan Inc, language model published 2026-06-29. Notable for: discretionary.

Solar Open2 250B from Upstage, language model published 2026-06-28. Notable for: training cost,discretionary.

Claude Mythos 5 from Anthropic, language,multimodal model published 2026-06-09. Notable for: sota improvement.

Claude Fable 5 from Anthropic, language,multimodal model published 2026-06-09. Notable for: sota improvement.

Nemotron 3 Ultra from Nvidia, language model published 2026-06-04. Notable for: discretionary.

Qwen3.7-Plus from Alibaba, language,vision,multimodal model published 2026-06-03. Notable for: discretionary.

MAI-Image-2.5 from Microsoft, image generation model published 2026-06-02. Notable for: discretionary.

MAI-Code-1-Flash from Microsoft, language model published 2026-06-02. Notable for: discretionary.

MiniMax-M3 from MiniMax, language,multimodal,vision,video model published 2026-06-01. Notable for: discretionary.

Claude Opus 4.8 from Anthropic, language model published 2026-05-28. Notable for: significant use.

Gemini 3.5 Flash from Google DeepMind, language,multimodal,audio,video,vision model published 2026-05-19. Notable for: significant use,discretionary.

Composer 2.5 from Cursor, language model published 2026-05-18. Notable for: discretionary.

TML-Interaction-Small from Thinking Machines, language,multimodal,audio,video model published 2026-05-11. Notable for: discretionary.

GPT-5.5 Cyber from OpenAI, cybersecurity model published 2026-05-07. Notable for: discretionary.

GPT-5.5 Instant from OpenAI, multimodal,language,vision model published 2026-05-05. Notable for: significant use.

DeepSeek-V4-Pro from DeepSeek, language model published 2026-04-24. Notable for: discretionary.

DeepSeek-V4-Flash from DeepSeek, language model published 2026-04-24. Notable for: discretionary.

Tencent Hy3 preview from Tencent, language model published 2026-04-23. Notable for: discretionary.

MiMo-V2.5-Pro from Xiaomi Corp, language model published 2026-04-23. Notable for: discretionary.

GPT-5.5 Pro from OpenAI, multimodal,language,vision model published 2026-04-23. Notable for: discretionary.

GPT-5.5 from OpenAI, multimodal,language,vision model published 2026-04-23. Notable for: significant use.

GPT Image 2 from OpenAI, image generation model published 2026-04-21. Notable for: discretionary.

Kimi K2.6 from Moonshot, language model published 2026-04-20. Notable for: discretionary.

Claude Opus 4.7 from Anthropic, language model published 2026-04-16. Notable for: discretionary.

Gemini Flash 3.1 TTS from Google DeepMind, audio model published 2026-04-15. Notable for: discretionary.

GPT-5.4-Cyber from OpenAI, cybersecurity,language model published 2026-04-14. Notable for: discretionary.

EXAONE 4.5 from LG AI Research, language,multimodal model published 2026-04-09. Notable for: training cost.

Muse Spark from Meta AI, multimodal,language,vision model published 2026-04-08. Notable for: discretionary.

GLM-5.1 from Z.ai (Zhipu AI), language model published 2026-04-07. Notable for: discretionary.

Claude Mythos Preview from Anthropic, language,multimodal model published 2026-04-07. Notable for: sota improvement.

MolmoAct 2 from Allen Institute for AI, University of Washington, National University of Singapore, University of Pennsylvania, Johns Hopkins University, Amazon, Cortex AI, vision,language,robotics model published 2026-04-05. Notable for: discretionary.

Qwen 3.6 Plus from Alibaba, language,multimodal model published 2026-04-01. Notable for: discretionary.

Qwen3.5-Omni-Plus from Alibaba, language,multimodal,vision,audio model published 2026-03-29. Notable for: discretionary.

Qwen3.5-Omni-Flash from Alibaba, language,multimodal,vision,audio model published 2026-03-29. Notable for: discretionary.

Composer 2 from Cursor, language model published 2026-03-19. Notable for: discretionary.

MiMo-V2-Pro from Xiaomi Corp, language model published 2026-03-18. Notable for: discretionary.

Nemotron 3 Super from Nvidia, language model published 2026-03-11. Notable for: discretionary.

GPT-5.4 Pro from OpenAI, multimodal,language,vision model published 2026-03-05. Notable for: discretionary.

GPT-5.4 from OpenAI, multimodal,language,vision model published 2026-03-05. Notable for: significant use.

Gemini 3.0 Flash-lite from Google DeepMind, language model published 2026-03-03. Notable for: significant use.

SWE 1.6 from Cognition, language model published 2026-03-01. Notable for: discretionary.

Qwen3.5-122B-A10B from Alibaba, language model published 2026-02-24. Notable for: discretionary.

Grok 4.20 from xAI, language model published 2026-02-17. Notable for: discretionary.

Qwen3.5 397B-A17B from Alibaba, language,vision model published 2026-02-15. Notable for: discretionary.

Doubao-Seed-2.0 Pro from ByteDance, language,vision,multimodal,video model published 2026-02-14. Notable for: significant use,discretionary.

Seedance 2.0 from ByteDance, image generation,video,audio model published 2026-02-12. Notable for: discretionary.

GPT-5.3-Codex-Spark from OpenAI, language model published 2026-02-12. Notable for: discretionary.

GLM-5 from Z.ai (Zhipu AI), language model published 2026-02-12. Notable for: discretionary.

GPT-5.3 Codex from OpenAI, language model published 2026-02-05. Notable for: significant use.

Qwen3-Coder-Next from Alibaba, language model published 2026-02-02. Notable for: discretionary.

Kimi K2.5 from Moonshot, language model published 2026-02-02. Notable for: discretionary.

Qwen3-Max-Thinking from Alibaba, language model published 2026-01-25. Notable for: discretionary.

Solar Open 100B from Upstage, language model published 2025-12-31. Notable for: training cost.

K-EXAONE from LG AI Research, language model published 2025-12-31. Notable for: training cost.

VAETKI from NC AI, language model published 2025-12-30. Notable for: training cost.

A.X K1 from SK Telecom, language model published 2025-12-30. Notable for: training cost.

HyperCLOVA X SEED 32B Think from NAVER, multimodal,language,vision model published 2025-12-29. Notable for: significant use.

MiniMax-M2.1 from MiniMax, language model published 2025-12-23. Notable for: discretionary.

GLM-4.7 from Z.ai (Zhipu AI), language model published 2025-12-22. Notable for: discretionary.

GPT-5.2 Codex from OpenAI, language model published 2025-12-18. Notable for: significant use.

Nemotron 3-Nano-30B-A3B from Nvidia, language model published 2025-12-15. Notable for: discretionary.

GPT-5.2 Pro from OpenAI, multimodal,language,vision model published 2025-12-11. Notable for: discretionary.

GPT-5.2 from OpenAI, published 2025-12-11. Notable for: significant use.

SIMA 2 from Google DeepMind, 3d modeling model published 2025-12-04. Notable for: discretionary.

DeepSeekMath-V2 from DeepSeek, language model published 2025-11-27. Notable for: sota improvement.

Claude Opus 4.5 from Anthropic, language,multimodal,vision model published 2025-11-24. Notable for: sota improvement.

Olmo 3 from Allen Institute for AI, language model published 2025-11-20. Notable for: sota improvement.

Gemini 3 Pro Image (Nano Banana Pro) from Google DeepMind, image generation model published 2025-11-20. Notable for: significant use.

Grok 4.1 Fast from xAI, language model published 2025-11-19. Notable for: sota improvement.

GPT-5.1-Codex-Max from OpenAI, language model published 2025-11-19. Notable for: significant use.

Gemini 3 Pro from Google DeepMind, multimodal,language,vision model published 2025-11-18. Notable for: significant use.

Gemini 3 Deep Think from Google DeepMind, language model published 2025-11-18. Notable for: discretionary.

π0.6 (pi-0.6) from Physical Intelligence, robotics,vision model published 2025-11-17. Notable for: discretionary.

P1-235B-A22B from Shanghai AI Lab, language model published 2025-11-17. Notable for: discretionary.

Grok 4.1 from xAI, language model published 2025-11-17. Notable for: sota improvement.

GPT-5.1 Instant from OpenAI, multimodal,language,vision model published 2025-11-13. Notable for: significant use.

GPT-5.1 from OpenAI, multimodal,language,vision model published 2025-11-13. Notable for: significant use.

GPT-5.1-Codex from OpenAI, language model published 2025-11-12. Notable for: significant use.

Meta's Generative Ads Model (GEM) from Meta AI, recommendation model published 2025-11-10. Notable for: significant use.

Kaiju Small from Character.ai, language model published 2025-11-07. Notable for: significant use.

Kaiju Medium from Character.ai, language model published 2025-11-07. Notable for: significant use.

Kaiju Large from Character.ai, language model published 2025-11-07. Notable for: significant use.

Kimi K2 Thinking from Moonshot, language model published 2025-11-06. Notable for: sota improvement.

Tongyi DeepResearch from Alibaba, language model published 2025-10-28. Notable for: sota improvement.

MiniMax-M2 from MiniMax, language model published 2025-10-27. Notable for: discretionary.

Veo 3.1 from Google DeepMind, video,vision model published 2025-10-15. Notable for: sota improvement.

Claude Haiku 4.5 from Anthropic, language model published 2025-10-15. Notable for: significant use.

Ling-1T from Ant Group, language model published 2025-10-10. Notable for: sota improvement.

GPT-5 Pro from OpenAI, multimodal,language,vision model published 2025-10-07. Notable for: significant use,historical significance.

Sora 2.0 from OpenAI, video model published 2025-09-30. Notable for: significant use.

GLM-4.6 from Z.ai (Zhipu AI), Tsinghua University, language model published 2025-09-30. Notable for: discretionary.

Claude Sonnet 4.5 from Anthropic, language,vision,multimodal model published 2025-09-29. Notable for: sota improvement.

Gemini Robotics-ER 1.5 from Google DeepMind, vision,language,speech model published 2025-09-25. Notable for: sota improvement.

Qwen3-Omni-30B-A3B from Alibaba, multimodal,language,vision,speech,video model published 2025-09-22. Notable for: sota improvement.

AgentFounder-30B from Alibaba, language model published 2025-09-16. Notable for: sota improvement.

Qwen3-Max from Alibaba, language model published 2025-09-05. Notable for: discretionary.

LongCat-Flash from Meituan Inc, language model published 2025-09-01. Notable for: sota improvement.

Gemini 2.5 Flash Image (Nano Banana) from Google, image generation model published 2025-08-26. Notable for: discretionary.

GPT-5 nano from OpenAI, multimodal,language,vision model published 2025-08-07. Notable for: significant use.

GPT-5 mini from OpenAI, multimodal,language,vision model published 2025-08-07. Notable for: significant use.

GPT-5 from OpenAI, multimodal,language,vision model published 2025-08-07. Notable for: significant use,historical significance.

gpt-oss-20b from OpenAI, language model published 2025-08-05. Notable for: discretionary.

OpenAI's first open-weights model since GPT-2, August 2025. Apache 2.0.

Claude Opus 4.1 from Anthropic, language,multimodal,vision model published 2025-08-05. Notable for: training cost.

Qwen Image from Alibaba, image generation,vision model published 2025-08-04. Notable for: sota improvement.

Hierarchical Reasoning Model (HPM) from Sapient Intelligence, language,vision,multimodal model published 2025-08-04. Notable for: sota improvement.

MindLink-72B from Kunlun Inc., language model published 2025-08-01. Notable for: sota improvement.

Gemini 2.5 Deep Think from Google, Google DeepMind, language,multimodal,vision,video,audio,mathematics model published 2025-08-01. Notable for: sota improvement.

GLM-4.5 from Z.ai (Zhipu AI), Tsinghua University, language model published 2025-07-28. Notable for: discretionary.

Qwen3-235B-A22B-Thinking (Jul 2025) from Alibaba, language model published 2025-07-25. Notable for: discretionary.

Qwen3-235B-A22B (Jul 2025) from Alibaba, language model published 2025-07-25. Notable for: discretionary.

Qwen3-Coder-480B-A35B from Alibaba, language model published 2025-07-22. Notable for: discretionary.

EXAONE 4.0 (32B) from LG AI Research, language model published 2025-07-15. Notable for: training cost.

Gemini Embedding from Google DeepMind, language model published 2025-07-14. Notable for: sota improvement.

Kimi K2 from Moonshot, language model published 2025-07-11. Notable for: training cost.

Grok 4 Heavy from xAI, language model published 2025-07-10. Notable for: sota improvement.

Grok 4 from xAI, language,multimodal,vision model published 2025-07-09. Notable for: training cost.

EXAONE Path 2.0 from LG AI Research, vision,medicine model published 2025-07-09. Notable for: sota improvement.

FGN from Google DeepMind, earth science model published 2025-06-12. Notable for: sota improvement.

Seed-1.6-Thinking from ByteDance, multimodal,language,vision model published 2025-06-11. Notable for: sota improvement.

Qwen3 Embedding from Alibaba, language model published 2025-06-05. Notable for: sota improvement.

Gemini 2.5 Pro (Jun 2025) from Google DeepMind, language,vision,video,multimodal,speech model published 2025-06-05. Notable for: sota improvement.

DeepSeek-R1 (May 2025) from DeepSeek, language model published 2025-05-28. Notable for: training cost,sota improvement.

Claude Sonnet 4 from Anthropic, language,multimodal,vision model published 2025-05-22. Notable for: training cost.

Anthropic's May 2025 model, built for long agentic coding sessions.

Veo 3 from Google DeepMind, video,vision model published 2025-05-21. Notable for: sota improvement.

Imagen 4 from Google, image generation model published 2025-05-20. Notable for: significant use.

Seed1.5-VL from ByteDance, vision,language,multimodal,video model published 2025-05-11. Notable for: sota improvement.

Gemini 2.5 Pro (May 2025) from Google DeepMind, language,vision,video,multimodal,speech model published 2025-05-06. Notable for: sota improvement.

Qwen3-235B-A22B from Alibaba, language model published 2025-04-29. Notable for: discretionary.

gpt-image-1 from OpenAI, image generation,vision model published 2025-04-23. Notable for: significant use.

Pangu Ultra from Huawei, language model published 2025-04-10. Notable for: sota improvement.

Llama 4 Scout from Meta AI, multimodal,language,vision model published 2025-04-05. Notable for: discretionary.

Llama 4 Maverick from Meta AI, multimodal,language,vision model published 2025-04-05. Notable for: discretionary.

Llama 4 Behemoth (preview) from Meta AI, multimodal,language,vision model published 2025-04-05. Notable for: training cost.

GPT-4o (Mar 2025) from OpenAI, multimodal,language,audio,speech,vision model published 2025-03-27. Notable for: sota improvement,significant use.

Google's reasoning model with a million-token context, March 2025. Led Chatbot Arena for months.

DeepSeek-V3 (Mar 2025) from DeepSeek, language model published 2025-03-24. Notable for: training cost.

Diffusion Renderer from Nvidia, University of Toronto, Vector Institute, University of Illinois Urbana-Champaign (UIUC), video model published 2025-03-22. Notable for: sota improvement.

o1-pro from OpenAI, language,mathematics,multimodal model published 2025-03-19. Notable for: sota improvement,significant use.

EXAONE Deep 32B from LG AI Research, language model published 2025-03-16. Notable for: training cost.

ERNIE-4.5-VL-424B-A47B (文心大模型4.5) from Baidu, multimodal,language,vision,video model published 2025-03-16. Notable for: sota improvement.

Hunyuan-TurboS from Tencent, language model published 2025-03-11. Notable for: sota improvement.

QwQ-32B from Alibaba, language model published 2025-03-06. Notable for: sota improvement.

Mistral OCR from Mistral AI, multimodal,vision,language model published 2025-03-06. Notable for: sota improvement.

Mercury from Inception Labs, language model published 2025-02-27. Notable for: sota improvement.

GPT-4.5 from OpenAI, language,vision,multimodal model published 2025-02-27. Notable for: training cost.

Claude 3.7 Sonnet from Anthropic, language,vision,multimodal model published 2025-02-24. Notable for: training cost.

Grok 3 from xAI, language,vision,multimodal model published 2025-02-17. Notable for: training cost.

o3-mini from OpenAI, language model published 2025-01-31. Notable for: significant use.

GPT-4o (Jan 2025) from OpenAI, multimodal,language,audio,speech,vision model published 2025-01-29. Notable for: sota improvement,significant use.

Computer-Using Agent (CUA) from OpenAI, vision,language,multimodal model published 2025-01-23. Notable for: sota improvement.

Kimi k1.5 from Moonshot, multimodal,language,vision model published 2025-01-22. Notable for: sota improvement.

Doubao-1.5-pro from ByteDance, language model published 2025-01-22. Notable for: training cost.

Eagle 2 from Nvidia, Nanjing University, Tsinghua University, Hong Kong Polytechnic University, Johns Hopkins University, New York University (NYU), vision,robotics,language model published 2025-01-20. Notable for: sota improvement.

Open-weights reasoning model, MIT licence, January 2025. Matched o1 and moved markets.

INTELLECT-MATH from Prime Intellect, mathematics model published 2025-01-17. Notable for: sota improvement.

Stable Point Aware 3D (SPAR3D) from Stability AI, University of Illinois Urbana-Champaign (UIUC), 3d modeling model published 2025-01-08. Notable for: sota improvement.

Eurus-2-7B-PRIME from Tsinghua University, University of Illinois Urbana-Champaign (UIUC), Shanghai AI Lab, Peking University, Shanghai Jiao Tong University, CUHK Shenzhen Research Institute, mathematics,language model published 2025-01-02. Notable for: sota improvement.

STORM-B/8 from University of Southern California, Georgia Institute of Technology, Stanford University, Nvidia, 3d modeling model published 2024-12-31. Notable for: sota improvement.

671 billion parameters, 37 billion active. Trained for under $6 million of GPU time, by the paper's account.

o3 from OpenAI, language,vision,multimodal model published 2024-12-20. Notable for: significant use.

Veo 2 from Google DeepMind, video,vision model published 2024-12-16. Notable for: sota improvement.

Apollo 7B from Meta AI, Stanford University, multimodal,video,language model published 2024-12-13. Notable for: sota improvement.

Gemini 2.0 Pro from Google DeepMind, language,multimodal,vision,video,audio model published 2024-12-11. Notable for: training cost.

Sora Turbo from OpenAI, video,vision model published 2024-12-09. Notable for: significant use.

EXAONE 3.5 32B from LG AI Research, language model published 2024-12-09. Notable for: training cost.

Llama 3.3 70B from Meta AI, language model published 2024-12-06. Notable for: training cost.

OpenAI's first reasoning model. Preview in September 2024, full release in December.

NVILA 15B from Nvidia, Massachusetts Institute of Technology (MIT), University of California (UC) Berkeley, University of California San Diego, University of Washington, Tsinghua University, vision,language,multimodal,video model published 2024-12-05. Notable for: sota improvement.

Infinity from ByteDance, image generation model published 2024-12-05. Notable for: sota improvement.

Amazon Nova Pro from Amazon, multimodal,language,video,vision model published 2024-12-03. Notable for: sota improvement.

Fugatto 1 from Nvidia, multimodal,language,audio model published 2024-11-25. Notable for: sota improvement.

GPT-4o (Nov 2024) from OpenAI, multimodal,language,audio,speech,vision model published 2024-11-20. Notable for: sota improvement,significant use.

Suno v4 from Suno, audio model published 2024-11-19. Notable for: significant use.

Pixtral Large from Mistral AI, multimodal,language,vision model published 2024-11-18. Notable for: significant use,sota improvement.

k0-math from Moonshot, mathematics,language model published 2024-11-16. Notable for: sota improvement.

Gemini-Exp-1114 from Google DeepMind, language model published 2024-11-15. Notable for: training cost,significant use.

SeedEdit from ByteDance, vision model published 2024-11-11. Notable for: sota improvement.

Hunyuan-Large from Tencent, language model published 2024-11-06. Notable for: training cost.

Doubao-pro from ByteDance, language model published 2024-10-28. Notable for: training cost.

NVLM-X 72B from Nvidia, vision,language model published 2024-10-22. Notable for: training cost.

NVLM-H 72B from Nvidia, vision,language model published 2024-10-22. Notable for: training cost.

NVLM-D 72B from Nvidia, vision,language model published 2024-10-22. Notable for: sota improvement.

Yi-Lightning from 01.AI, language model published 2024-10-16. Notable for: training cost.

CHAI-1 from Chai discovery, biology model published 2024-10-15. Notable for: sota improvement.

Llama-3.1-Nemotron-70B-Instruct from Nvidia, Meta AI, language model published 2024-10-12. Notable for: sota improvement.

RDT-1B from Tsinghua University, robotics model published 2024-10-10. Notable for: sota improvement.

Palmyra X 004 from Writer, language model published 2024-10-09. Notable for: sota improvement.

GR-2 from ByteDance, robotics model published 2024-10-08. Notable for: sota improvement.

Movie Gen Video from Meta AI, video,vision model published 2024-10-04. Notable for: training cost.

PixelDance from ByteDance, video,vision model published 2024-09-24. Notable for: sota improvement.

Llama 3.2 11B from Meta AI, multimodal,vision,language model published 2024-09-24. Notable for: significant use.

Telechat2-115B from China Telecom, language model published 2024-09-20. Notable for: training cost.

Qwen2.5 Instruct (72B) from Alibaba, language model published 2024-09-19. Notable for: training cost.

Alibaba's open model that matched Llama 3.1 405B on many tasks at a fifth of the size.

Oryx 34B from Tsinghua University, Tencent, Nanyang Technological University, multimodal,vision,3d modeling,video,language model published 2024-09-19. Notable for: sota improvement.

Qwen2.5-32B from Alibaba, language model published 2024-09-17. Notable for: training cost.

o1-preview from OpenAI, language,mathematics,biology model published 2024-09-12. Notable for: sota improvement,significant use.

o1-mini from OpenAI, language model published 2024-09-12. Notable for: significant use.

DeepSeek-V2.5 from DeepSeek, language model published 2024-09-06. Notable for: training cost.

Hunyuan Turbo from Tencent, language model published 2024-09-05. Notable for: sota improvement.

Harrison.rad.1 from Harrison.ai, vision,medicine,language,multimodal model published 2024-09-05. Notable for: sota improvement.

AlphaProteo from Google DeepMind, biology model published 2024-09-05. Notable for: historical significance.

Hairuo from Inspur, language model published 2024-08-29. Notable for: sota improvement.

GLM-4-Plus from Z.ai (Zhipu AI), language model published 2024-08-29. Notable for: training cost.

Jamba 1.5-Large from AI21 Labs, language model published 2024-08-22. Notable for: training cost.

Grok-2 from xAI, language,vision,multimodal model published 2024-08-13. Notable for: training cost.

Table Tennis Agent from Google DeepMind, robotics model published 2024-08-07. Notable for: sota improvement.

LLaVA-OV-72B from ByteDance, Nanyang Technological University, Chinese University of Hong Kong (CUHK), Hong Kong University of Science and Technology (HKUST), multimodal,vision,language,video model published 2024-08-06. Notable for: sota improvement.

GPT-4o (Aug 2024) from OpenAI, multimodal,language,audio,speech,vision model published 2024-08-06. Notable for: sota improvement,significant use.

AFM-server from Apple, language model published 2024-07-29. Notable for: significant use.

AFM-on-device from Apple, language model published 2024-07-29. Notable for: significant use.

Mistral Large 2 from Mistral AI, language model published 2024-07-24. Notable for: training cost.

The first open-weights model at GPT-4 level, July 2024. 15 trillion training tokens.

GPT-4o mini from OpenAI, language,multimodal,vision model published 2024-07-18. Notable for: significant use.

Mathstral from Mistral AI, mathematics,language model published 2024-07-16. Notable for: sota improvement.

DeepL LLM from DeepL, language model published 2024-07-16. Notable for: significant use.

SenseChat 5.5 from SenseTime, multimodal,language,vision model published 2024-07-06. Notable for: sota improvement.

Ernie 4.0 Turbo from Baidu, multimodal,language,vision model published 2024-06-28. Notable for: significant use.

ESM3 (98B) from EvolutionaryScale, University of California (UC) Berkeley, biology model published 2024-06-25. Notable for: historical significance.

Cambrian-1-34B from New York University (NYU), multimodal,vision,language model published 2024-06-24. Notable for: sota improvement.

The model that made Claude the coder's default in 2024. Computer use arrived with the October update.

DeepSeek-Coder-V2 236B from DeepSeek, language model published 2024-06-17. Notable for: sota improvement.

Nemotron-4 340B from Nvidia, language model published 2024-06-14. Notable for: training cost.

OpenVLA from Stanford University, University of California (UC) Berkeley, Toyota Research Institute, Google DeepMind, Massachusetts Institute of Technology (MIT), Physical Intelligence, robotics,vision,language model published 2024-06-13. Notable for: sota improvement.

Qwen2-72B from Alibaba, language model published 2024-06-07. Notable for: training cost.

ALLaM adapted 70B from Saudi Data and Artificial Intelligence Authority, language model published 2024-05-21. Notable for: sota improvement.

Octo-Base from University of California (UC) Berkeley, Stanford University, Carnegie Mellon University (CMU), DeepMind, robotics model published 2024-05-20. Notable for: sota improvement.

GLM-4 (0520) from Z.ai (Zhipu AI), language model published 2024-05-20. Notable for: training cost.

Yi-Large from 01.AI, language model published 2024-05-13. Notable for: training cost.

Omni: text, image and audio in one model with real-time voice. May 2024.

AlphaFold 3 from Google DeepMind, Isomorphic Labs, biology model published 2024-05-08. Notable for: sota improvement.

VILA1.5-13B from Nvidia, Massachusetts Institute of Technology (MIT), multimodal,language,vision,video model published 2024-05-03. Notable for: sota improvement.

GenCast from Google DeepMind, earth science model published 2024-05-01. Notable for: sota improvement.

Llama 3-70B from Meta AI, language model published 2024-04-18. Notable for: significant use.

Reka Core from Reka AI, multimodal,language,vision,video,speech model published 2024-04-15. Notable for: training cost.

GPT-4 Turbo (Apr 2024) from OpenAI, multimodal,vision,language,image generation model published 2024-04-09. Notable for: sota improvement.

ReALM from Apple, language model published 2024-03-29. Notable for: sota improvement.

DBRX from Databricks, language model published 2024-03-27. Notable for: training cost.

MM1-30B from Apple, multimodal,language,vision model published 2024-03-14. Notable for: sota improvement.

ManiGaussian from Tsinghua University, Nanyang Technological University, Carnegie Mellon University (CMU), robotics,vision,video model published 2024-03-13. Notable for: sota improvement.

Inflection-2.5 from Inflection AI, language model published 2024-03-07. Notable for: significant use.

Claude 3 Sonnet from Anthropic, multimodal,language,vision model published 2024-03-04. Notable for: training cost.

Anthropic's first model to match GPT-4, March 2024.

Aramco Metabrain AI from Saudi Aramco, language model published 2024-03-04. Notable for: training cost.

Mistral Large from Mistral AI, language model published 2024-02-26. Notable for: training cost.

MegaScale (Production) from ByteDance, Peking University, language model published 2024-02-23. Notable for: sota improvement.

Stable Diffusion 3 from Stability AI, image generation model published 2024-02-22. Notable for: sota improvement.

Sora from OpenAI, video,vision model published 2024-02-15. Notable for: sota improvement.

Gemini 1.5 Pro from Google DeepMind, language,multimodal model published 2024-02-15. Notable for: significant use.

Aya from Cohere Labs (formerly Cohere for AI), Brown University, Cohere, Carnegie Mellon University (CMU), Massachusetts Institute of Technology (MIT), language model published 2024-02-12. Notable for: sota improvement.

Qwen1.5-72B from Alibaba, language model published 2024-02-04. Notable for: sota improvement.

Qwen-VL-Max from Alibaba, multimodal,language,vision model published 2024-01-25. Notable for: sota improvement.

AlphaGeometry from Google DeepMind, New York University (NYU), mathematics model published 2024-01-17. Notable for: sota improvement.

Palmyra X 003 from Writer, language model published 2024-01-01. Notable for: sota improvement.

Kimi Explorer from Moonshot, language model published 2024-01-01. Notable for: sota improvement.

CoRe from Tsinghua University, mathematics,language model published 2023-12-29. Notable for: sota improvement.

GQA-8-XXL from Google Research, language model published 2023-12-23. Notable for: historical significance.

nekomata-14b from rinna, language model published 2023-12-21. Notable for: sota improvement.

Gemini Nano-2 from Google DeepMind, multimodal,language,vision,audio model published 2023-12-19. Notable for: significant use.

Gemini Nano-1 from Google DeepMind, multimodal,language,vision,audio model published 2023-12-19. Notable for: significant use.

FunSearch from Google DeepMind, language,search model published 2023-12-14. Notable for: sota improvement,historical significance.

CogAgent from Tsinghua University, Z.ai (Zhipu AI), vision,language model published 2023-12-14. Notable for: sota improvement.

VILA-13B from Nvidia, Massachusetts Institute of Technology (MIT), multimodal,language,vision model published 2023-12-12. Notable for: sota improvement.

W.A.L.T from Stanford University, Google Research, Georgia Institute of Technology, video model published 2023-12-11. Notable for: sota improvement.

Eight experts, two active per token. GPT-3.5 quality with open weights.

SeamlessM4T from Facebook, INRIA, University of California (UC) Berkeley, speech,language model published 2023-12-08. Notable for: sota improvement.

Llama Guard from Meta AI, language model published 2023-12-07. Notable for: sota improvement.

Google DeepMind's first frontier Gemini, December 2023. The first model to pass 90% on MMLU.

Gemini 1.0 Pro from Google DeepMind, multimodal,language,vision model published 2023-12-06. Notable for: significant use.

Mamba-24M (SC09) from Carnegie Mellon University (CMU), Princeton University, speech model published 2023-12-01. Notable for: sota improvement.

Qwen-72B from Alibaba, language model published 2023-11-30. Notable for: sota improvement.

PPLX-70B-Online from Perplexity, language model published 2023-11-29. Notable for: significant use.

GNoME for crystal discovery from Google DeepMind, materials science model published 2023-11-29. Notable for: historical significance.

Inflection-2 from Inflection AI, language model published 2023-11-22. Notable for: significant use,training cost.

Claude 2.1 from Anthropic, language model published 2023-11-21. Notable for: significant use.

AndesGPT from Oppo Mobile Telecommunications, language model published 2023-11-16. Notable for: sota improvement.

Nemotron-3-8B from Nvidia, language model published 2023-11-15. Notable for: sota improvement.

Qwen-Audio-Chat from Alibaba, language,speech,audio model published 2023-11-14. Notable for: sota improvement.

GraphCast from Google DeepMind, earth science model published 2023-11-14. Notable for: sota improvement.

Volcano 13B from Korea University, Korea Advanced Institute of Science and Technology (KAIST), LG, language,multimodal,vision model published 2023-11-13. Notable for: sota improvement.

SPHINX (Llama 2 13B) from Shanghai AI Lab, Chinese University of Hong Kong (CUHK), ShanghaiTech University, vision,language,multimodal model published 2023-11-13. Notable for: sota improvement.

RoFormer from Zhuiyi Technology, language model published 2023-11-08. Notable for: historical significance.

MultiBand Diffusion from Meta AI, Hebrew University of Jerusalem, LORIA, audio,speech model published 2023-11-08. Notable for: sota improvement.

OmniVec from TensorTour, multimodal,vision,speech,language,video,3d modeling model published 2023-11-07. Notable for: sota improvement.

mPLUG-Owl2 from Alibaba, vision,language,multimodal model published 2023-11-07. Notable for: sota improvement.

GPT-4 Turbo (Nov 2023) from OpenAI, multimodal,vision,language,image generation model published 2023-11-06. Notable for: sota improvement.

CogVLM-17B from Tsinghua University, Z.ai (Zhipu AI), Beihang University, multimodal,vision,language model published 2023-11-06. Notable for: sota improvement.

LLaVA 1.5 from University of Wisconsin Madison, Microsoft Research, multimodal,language,vision model published 2023-11-05. Notable for: sota improvement.

Grok-1 from xAI, language model published 2023-11-04. Notable for: training cost.

RT-Trajectory from Google DeepMind, University of California San Diego, Stanford University, robotics model published 2023-11-03. Notable for: sota improvement.

BLUUMI from University of Turku, Hugging Face, language model published 2023-11-03. Notable for: sota improvement.

Yi-34B from 01.AI, language model published 2023-11-02. Notable for: significant use.

Cohere Embed from Cohere, language model published 2023-11-02. Notable for: sota improvement.

Skywork-13B from Kunlun Inc., language model published 2023-10-30. Notable for: sota improvement.

ChatGLM3-6B from Z.ai (Zhipu AI), multimodal,language,vision model published 2023-10-27. Notable for: sota improvement.

DiT-XL/2 + CADS from ETH Zurich, Disney Research, image generation model published 2023-10-26. Notable for: sota improvement.

CODEFUSION (Python) from Microsoft, Microsoft Research, language model published 2023-10-26. Notable for: sota improvement.

DALL·E 3 from OpenAI, image generation model published 2023-10-19. Notable for: sota improvement.

PaLI-3 from Google DeepMind, Google Research, Google Cloud, multimodal,language,vision model published 2023-10-17. Notable for: sota improvement.

ERNIE 4.0 from Baidu, multimodal,language,video,image generation model published 2023-10-17. Notable for: significant use.

RT-2-X from Google DeepMind, robotics model published 2023-10-13. Notable for: sota improvement.

Ferret (13B) from Columbia University, Apple, multimodal,language,vision model published 2023-10-11. Notable for: sota improvement.

RoseTTAFold All-Atom (RFAA) from University of Washington, Seoul National University, University of Sheffield, biology model published 2023-10-09. Notable for: sota improvement.

FinGPT-13B from University of California Los Angeles (UCLA), Columbia University, New York University (NYU), language model published 2023-10-07. Notable for: sota improvement.

CTM (CIFAR-10) from Stanford University, Sony, image generation model published 2023-10-01. Notable for: sota improvement.

GPT-3.5 Turbo Instruct from OpenAI, language model published 2023-09-28. Notable for: historical significance,significant use.

Amazon Titan from Amazon, language,image generation model published 2023-09-28. Notable for: training cost.

Show-1 from National University of Singapore, video model published 2023-09-27. Notable for: sota improvement.

Mistral AI's first model, Apache 2.0, better than Llama 2 13B at half the size.

GPT-4V from OpenAI, multimodal,vision,language model published 2023-09-25. Notable for: significant use.

AlphaMissense from Google DeepMind, biology model published 2023-09-22. Notable for: sota improvement.

Robot Parkour from Shanghai Qi Zhi institute, Stanford University, Carnegie Mellon University (CMU), Tsinghua University, robotics model published 2023-09-12. Notable for: sota improvement.

Falcon-180B from Technology Innovation Institute, language model published 2023-09-06. Notable for: training cost.

Swift from Intel Labs, robotics model published 2023-08-30. Notable for: sota improvement.

Jais from Cerebras Systems, Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), Inception G42, language model published 2023-08-29. Notable for: sota improvement.

PeptideBERT from Carnegie Mellon University (CMU), biology model published 2023-08-28. Notable for: sota improvement.

Qwen-VL from Alibaba, multimodal,language,vision model published 2023-08-24. Notable for: sota improvement.

GGNN from Westlake University, Tsinghua University, Toyota Technological Institute at Chicago, biology model published 2023-08-05. Notable for: sota improvement.

RT-2 from Google DeepMind, robotics model published 2023-07-28. Notable for: sota improvement.

AudioLM from Google Research, audio model published 2023-07-26. Notable for: sota improvement.

Llama 2-7B from Meta AI, language model published 2023-07-18. Notable for: historical significance,significant use,highly cited.

The first Llama licensed for commercial use. The base of thousands of fine-tunes.

GPT3-2.7B (FlashAttention-2) from Stanford University, Princeton University, language model published 2023-07-18. Notable for: highly cited,historical significance.

Claude 2 from Anthropic, language model published 2023-07-11. Notable for: historical significance.

xTrimoPGLM -100B from Tsinghua University, BioMap Research, biology model published 2023-07-06. Notable for: sota improvement,training cost.

InternLM from Shanghai AI Lab, SenseTime, language model published 2023-07-06. Notable for: sota improvement.

Pangu-Weather from Huawei, earth science model published 2023-07-05. Notable for: sota improvement.

Stable Diffusion XL (SDXL) from Stability AI, image generation model published 2023-07-04. Notable for: significant use.

HyenaDNA from Stanford University, Harvard University, Mila - Quebec AI (originally Montreal Institute for Learning Algorithms), University of Montreal / Université de Montréal, biology model published 2023-06-27. Notable for: sota improvement.

ERNIE 3.5 from Baidu, language model published 2023-06-27. Notable for: sota improvement.

GPT-4 (Jun 2023) from OpenAI, multimodal,language,vision model published 2023-06-13. Notable for: highly cited,sota improvement,training cost.

The model behind the ChatGPT launch and the first chat API, March 2023.

MusicGen from Meta AI, audio model published 2023-06-08. Notable for: sota improvement.

LTM-1 from Magic, language model published 2023-06-06. Notable for: sota improvement.

PaLI-X from Google Research, multimodal,language,vision,video model published 2023-05-29. Notable for: sota improvement.

DPO on Pythia-2.8B from Stanford University, CZ Biohub Network, language model published 2023-05-29. Notable for: historical significance.

Goat-7B from National University of Singapore, language model published 2023-05-23. Notable for: sota improvement.

CodeT5+ from Salesforce, language model published 2023-05-20. Notable for: sota improvement.

ONE-PEACE from Alibaba, Huazhong University of Science and Technology, multimodal,vision,speech,language model published 2023-05-18. Notable for: sota improvement.

CoEdiT-xxl from University of Minnesota, Grammarly, language model published 2023-05-17. Notable for: sota improvement.

Med-PaLM 2 from Google Research, DeepMind, medicine,language model published 2023-05-16. Notable for: sota improvement.

InstructBLIP from Salesforce Research, Hong Kong University of Science and Technology (HKUST), Nanyang Technological University, multimodal,language,vision model published 2023-05-11. Notable for: sota improvement.

PaLM 2 from Google, language model published 2023-05-10. Notable for: sota improvement,training cost,significant use,highly cited.

StarCoder from Hugging Face, ServiceNow, Northeastern University, Mila - Quebec AI (originally Montreal Institute for Learning Algorithms), Carnegie Mellon University (CMU), Johns Hopkins University, Leipzig University, ScaDS.AI, Queen Mary University of London, Roblox, Sea AI Lab, Technion - Israel Institute of Technology, Monash University, CSIRO, Data61, McGill University, Saama, University of British Columbia (UBC), Massachusetts Institute of Technology (MIT), Technical University of Munich, IBM, University of Vermont, UnfoldML, SAP, University of Notre Dame, Columbia University, New York University (NYU), University of Allahabad, Discover Dollar, Toloka, Telefonica, Stanford University, Weizmann Institute of Science, Alan Turing Institute, Wellesley College, EleutherAI, Forschungszentrum Julich, language model published 2023-05-09. Notable for: sota improvement.

ImageBind from Meta AI, multimodal,vision,audio,language,image generation,speech model published 2023-05-09. Notable for: sota improvement.

Agile Soccer Robot from Google DeepMind, robotics model published 2023-04-26. Notable for: sota improvement.

LLaVA from University of Wisconsin Madison, Microsoft Research, Columbia University, multimodal,vision,language model published 2023-04-17. Notable for: sota improvement.

DINOv2 from Facebook AI Research, INRIA, vision model published 2023-04-14. Notable for: sota improvement.

Incoder-6.7B from Facebook AI Research, University of Washington, University of California (UC) Berkeley, Carnegie Mellon University (CMU), Toyota Technological Institute at Chicago, language model published 2023-04-09. Notable for: sota improvement.

Segment Anything Model from Meta AI, vision model published 2023-04-05. Notable for: highly cited,historical significance.

BloombergGPT from Bloomberg, Johns Hopkins University, language model published 2023-03-30. Notable for: sota improvement.

VideoMAE V2 from Nanjing University, Shenzhen Institute of Advanced Technology, Shanghai AI Lab, video model published 2023-03-29. Notable for: sota improvement.

SigLIP 400M from Google DeepMind, vision model published 2023-03-27. Notable for: significant use.

Firefly from Adobe, image generation model published 2023-03-21. Notable for: significant use.

PanGu-Σ from Huawei Noah's Ark Lab, language model published 2023-03-20. Notable for: sota improvement.

Gen-2 from Runway, video model published 2023-03-20. Notable for: sota improvement.

LEP-AD from King Abdullah University of Science and Technology (KAUST), Karolinska Institute, biology model published 2023-03-15. Notable for: sota improvement.

OpenAI's frontier model of 2023 and 2024. Reported as a mixture of experts. The first with a closed technical report.

Falcon-40B from Technology Innovation Institute, language model published 2023-03-15. Notable for: historical significance.

Claude from Anthropic, language model published 2023-03-14. Notable for: historical significance,sota improvement.

PaLM-E from Google, TU Berlin, robotics,vision,language model published 2023-03-06. Notable for: sota improvement.

AudioGen from Meta AI, Hebrew University of Jerusalem, audio model published 2023-03-05. Notable for: sota improvement.

DiT-XL/2 from New York University (NYU), University of California (UC) Berkeley, image generation model published 2023-03-02. Notable for: sota improvement.

Meta's 65 billion parameter model trained only on public data. The leak made it the base of the open ecosystem.

BASIC-L + Lion from Google, University of California Los Angeles (UCLA), vision model published 2023-02-13. Notable for: sota improvement.

ViT-22B from Google, vision model published 2023-02-10. Notable for: sota improvement.

ProteinDT from University of California (UC) Berkeley, California Institute of Technology, University of Toronto, University of Wisconsin Madison, Texas A&M, Nvidia, Mila - Quebec AI (originally Montreal Institute for Learning Algorithms), biology,language model published 2023-02-09. Notable for: sota improvement.

Gen-1 from Runway, video model published 2023-02-06. Notable for: sota improvement.

BLIP-2 (Q-Former) from Salesforce Research, vision,language model published 2023-01-30. Notable for: sota improvement.

DDPM-IP (CelebA) from Utrecht University, image generation model published 2023-01-27. Notable for: sota improvement.

MusicLM from Google, audio model published 2023-01-26. Notable for: sota improvement.

Ankh_large from Technical University of Munich, Columbia University, biology model published 2023-01-16. Notable for: sota improvement.

Nucleotide Transformer from Nvidia, Technical University of Munich, InstaDeep, biology model published 2023-01-15. Notable for: sota improvement.

DreamerV3 from DeepMind, University of Toronto, games model published 2023-01-10. Notable for: sota improvement.

VALL-E from Microsoft, audio,speech model published 2023-01-05. Notable for: sota improvement.

Hybrid H3-2.7B from Stanford University, University at Buffalo, language model published 2022-12-28. Notable for: sota improvement.

CaLM from University of Oxford, biology model published 2022-12-19. Notable for: sota improvement.

RT-1 from Google, robotics model published 2022-12-13. Notable for: sota improvement.

TranceptEve from University of Oxford, Harvard Medical School, biology model published 2022-12-10. Notable for: sota improvement.

Vega v2 from Wuhan University, JD Explore Academy, Shanghai AI Lab, Nanyang Technological University, Washington University in St Louis, Chongqing University of Posts and Telecommunications, University of Sydney, language model published 2022-12-04. Notable for: sota improvement.

DeepNash from DeepMind, games model published 2022-12-01. Notable for: sota improvement.

DiT-XL/2 + Discriminator Guidance from Korea Advanced Institute of Science and Technology (KAIST), NAVER, image generation model published 2022-11-28. Notable for: sota improvement.

Discriminator Guidance from Korea Advanced Institute of Science and Technology (KAIST), NAVER, image generation model published 2022-11-28. Notable for: sota improvement.

ALM 1.0 from Beijing Academy of Artificial Intelligence / BAAI, language model published 2022-11-28. Notable for: sota improvement.

CICERO from Meta AI, games model published 2022-11-22. Notable for: sota improvement.

AR-LDM from Alibaba, University of Waterloo, Vector Institute, image generation model published 2022-11-20. Notable for: sota improvement.

Fusion in Encoder from Samsung, language model published 2022-11-18. Notable for: sota improvement.

Galactica from Meta AI, language,biology model published 2022-11-16. Notable for: sota improvement.

EVA-01 from Beijing Academy of Artificial Intelligence / BAAI, Huazhong University of Science and Technology, Zhejiang University (ZJU), Beijing Institute of Technology, vision model published 2022-11-14. Notable for: sota improvement.

AltCLIP_M9 from Beijing Academy of Artificial Intelligence / BAAI, multimodal,language,vision model published 2022-11-12. Notable for: sota improvement.

InternImage from Shanghai AI Lab, Tsinghua University, Nanjing University, SenseTime, Chinese University of Hong Kong (CUHK), vision model published 2022-11-10. Notable for: sota improvement.

mT0-13B from Hugging Face, BigScience, language model published 2022-11-03. Notable for: sota improvement.

Mogrifier RLSTM (WT2) from DeepMind, language model published 2022-11-03. Notable for: sota improvement.

eDiff-I from Nvidia, image generation model published 2022-11-02. Notable for: sota improvement.

EnCodec from Meta AI, audio model published 2022-10-24. Notable for: sota improvement.

U-PaLM (540B) from Google, language model published 2022-10-20. Notable for: sota improvement.

LMSI-Palm from Google, University of Illinois Urbana-Champaign (UIUC), language model published 2022-10-20. Notable for: sota improvement.

Flan-PaLM 540B from Google, language model published 2022-10-20. Notable for: highly cited,sota improvement.

GenSLM from University of Chicago, Nvidia, Harvard University, Cerebras Systems, Technical University of Munich, California Institute of Technology, biology model published 2022-10-11. Notable for: sota improvement.

Diplodocus from Meta AI, Massachusetts Institute of Technology (MIT), games model published 2022-10-11. Notable for: sota improvement.

Phenaki from University College London (UCL), University of Michigan, Google Brain, video model published 2022-10-05. Notable for: sota improvement.

AlphaTensor from DeepMind, other,games,mathematics model published 2022-10-05. Notable for: sota improvement.

DiffDock from Massachusetts Institute of Technology (MIT), biology model published 2022-10-04. Notable for: sota improvement.

Make-A-Video from Meta AI, video model published 2022-09-29. Notable for: sota improvement.

Open speech recognition in 99 languages.

SauTech from Saudi Data and Artificial Intelligence Authority, Saudi Company for Artificial Intelligence, speech,audio model published 2022-09-14. Notable for: sota improvement.

PaLI from Google, language,vision,multimodal model published 2022-09-14. Notable for: sota improvement.

BEIT-3 from Microsoft, multimodal,vision,language model published 2022-08-22. Notable for: sota improvement.

BlenderBot 3 from McGill University, Meta AI, Mila - Quebec AI (originally Montreal Institute for Learning Algorithms), language model published 2022-08-10. Notable for: sota improvement.

GLM-130B from Tsinghua University, language model published 2022-08-04. Notable for: sota improvement.

AlexaTM 20B from Amazon, language model published 2022-08-02. Notable for: sota improvement.

OmegaPLM from Massachusetts Institute of Technology (MIT), Westlake University, biology model published 2022-07-22. Notable for: historical significance.

ESM2-15B from Meta AI, New York University (NYU), Stanford University, Massachusetts Institute of Technology (MIT), biology model published 2022-07-21. Notable for: sota improvement.

BLOOM-176B from Hugging Face, BigScience, language model published 2022-07-11. Notable for: historical significance,highly cited.

NLLB from Meta AI, language model published 2022-07-06. Notable for: sota improvement.

CodeT5-large from Salesforce, language model published 2022-07-05. Notable for: sota improvement.

Minerva (540B) from Google, language model published 2022-06-29. Notable for: sota improvement.

ProGen2-xlarge from Salesforce Research, Columbia University, Johns Hopkins University, biology model published 2022-06-27. Notable for: sota improvement.

Parti from Google Research, image generation model published 2022-06-22. Notable for: sota improvement.

CoCa from Google Research, vision model published 2022-06-14. Notable for: sota improvement.

MetaLM from Microsoft Research, multimodal,language,vision model published 2022-06-13. Notable for: sota improvement.

DITTO from Tsinghua University, Apple, Westlake University, Chinese University of Hong Kong (CUHK), language model published 2022-06-06. Notable for: sota improvement.

Diffusion-GAN from UT Austin, Microsoft, image generation model published 2022-06-05. Notable for: sota improvement.

CogVideo from Tsinghua University, Beijing Academy of Artificial Intelligence / BAAI, video model published 2022-05-29. Notable for: historical significance.

Tranception from University of Oxford, Harvard Medical School, Cohere, biology model published 2022-05-27. Notable for: sota improvement.

GPT-2 Medium (FlashAttention) from Stanford University, University at Buffalo, language model published 2022-05-27. Notable for: highly cited,historical significance.

Imagen from Google Brain, image generation model published 2022-05-23. Notable for: significant use,sota improvement,highly cited.

SimCSE from Princeton University, Tsinghua University, language model published 2022-05-18. Notable for: highly cited,sota improvement.

Gato from DeepMind, multimodal,robotics,games,language model published 2022-05-12. Notable for: sota improvement.

UL2 from Google Research, Google Brain, language model published 2022-05-10. Notable for: sota improvement.

DeBERTaV3large + KEAR from Microsoft, language model published 2022-05-04. Notable for: sota improvement.

OPT-175B from Meta AI, language model published 2022-05-02. Notable for: significant use,highly cited.

Flamingo from DeepMind, multimodal,vision,language,video model published 2022-04-29. Notable for: sota improvement,discretionary.

XMC-GAN from Google Research, image generation model published 2022-04-14. Notable for: sota improvement.

Sparse all-MLP from Meta AI, language model published 2022-04-14. Notable for: sota improvement.

Latent diffusion with a CLIP text encoder, trained on LAION. Open weights, runs on a consumer GPU.

BERT-RBP from Waseda University, biology model published 2022-04-07. Notable for: sota improvement.

DALL·E 2 from OpenAI, image generation model published 2022-04-06. Notable for: highly cited,sota improvement.

Google's 540 billion parameter model on 6,144 TPU v4 chips.

70 billion parameters on 1.4 trillion tokens. Beat Gopher with a quarter of the parameters and the same compute.

Make-A-Scene from Meta AI, image generation model published 2022-03-24. Notable for: sota improvement.

Segatron-XL large, M=384 + HCP from Microsoft Research, University of Waterloo, language model published 2022-03-21. Notable for: sota improvement.

GPT-3.5 (davinci-002) from OpenAI, language model published 2022-03-15. Notable for: historical significance,significant use,sota improvement,training cost.

ViT-G (model soup) from University of Washington, Columbia University, Google, Meta AI, Tel Aviv University, vision model published 2022-03-10. Notable for: sota improvement.

MegaSyn from Collaborations Pharmaceuticals, medicine model published 2022-03-07. Notable for: historical significance.

Statement Curriculum Learning from OpenAI, language,mathematics model published 2022-03-02. Notable for: sota improvement.

DeepNet from Microsoft Research, language model published 2022-03-01. Notable for: sota improvement.

PolyCoder from Carnegie Mellon University (CMU), language model published 2022-02-26. Notable for: sota improvement.

FourCastNet from Nvidia, NERSC, Lawrence Berkeley National Laboratory, University of Michigan, Rice University, California Institute of Technology, Purdue University, earth science model published 2022-02-22. Notable for: sota improvement.

ST-MoE from Google, Google Brain, Google Research, language model published 2022-02-17. Notable for: sota improvement.

Midjourney V1 from Midjourney, image generation model published 2022-02-15. Notable for: significant use,historical significance.

ProteinBERT from Hebrew University of Jerusalem, Ben-Gurion University of the Negev, Deep Trading, biology model published 2022-02-10. Notable for: sota improvement.

LaMDA from Google, language model published 2022-02-10. Notable for: historical significance.

GPT-NeoX-20B from EleutherAI, language model published 2022-02-09. Notable for: historical significance.

MaskGIT (ImageNet) from Google Research, image generation model published 2022-02-08. Notable for: sota improvement.

RETRO-7B from DeepMind, language model published 2022-02-07. Notable for: sota improvement.

AlphaCode from DeepMind, language model published 2022-02-02. Notable for: sota improvement.

InstructGPT 6B from OpenAI, language model published 2022-01-27. Notable for: historical significance,highly cited.

InstructGPT 1.3B from OpenAI, language model published 2022-01-27. Notable for: historical significance,highly cited.

GPT-3 taught to follow instructions with RLHF. The direct ancestor of ChatGPT.

OntoProtein from Zhejiang University (ZJU), biology,language model published 2022-01-23. Notable for: sota improvement.

AbLang (heavy sequences) from University of Oxford, biology model published 2022-01-22. Notable for: sota improvement.

data2vec (vision) from Meta AI, vision model published 2022-01-20. Notable for: sota improvement.

data2vec (speech) from Meta AI, speech model published 2022-01-20. Notable for: sota improvement.

data2vec (language) from Meta AI, language model published 2022-01-20. Notable for: sota improvement.

Detic from Meta AI, University of Texas at Austin, vision model published 2022-01-07. Notable for: sota improvement.

ERNIE-ViLG from Baidu, multimodal,image generation,vision,language model published 2021-12-31. Notable for: sota improvement.

ERNIE 3.0 Titan from Baidu, Peng Cheng Laboratory, language model published 2021-12-23. Notable for: sota improvement.

XGLM-7.5B from Meta AI, Facebook AI Research, language model published 2021-12-20. Notable for: sota improvement.

LDM-1.45B from Heidelberg University, Runway, image generation model published 2021-12-20. Notable for: highly cited.

Contriever from Meta AI, University College London (UCL), PSL University, Université Grenoble Alpes, language model published 2021-12-16. Notable for: sota improvement.

LongT5 from Google Research, language model published 2021-12-15. Notable for: sota improvement.

GLaM from Google, language model published 2021-12-13. Notable for: sota improvement.

DeepMind's 280 billion parameter model, December 2021.

Student of Games from DeepMind, games model published 2021-12-06. Notable for: sota improvement.

T-NLRv5 XXL from Microsoft, language model published 2021-12-03. Notable for: sota improvement.

NÜWA from Microsoft Research, Peking University, multimodal,vision,image generation,video,language model published 2021-11-24. Notable for: sota improvement.

Florence from Microsoft, vision model published 2021-11-22. Notable for: historical significance,sota improvement.

BASIC-L from Google, vision model published 2021-11-19. Notable for: sota improvement.

Swin Transformer V2 (SwinV2-G) from Microsoft Research Asia, vision,video model published 2021-11-18. Notable for: sota improvement,highly cited.

AliceMind-MMU from Alibaba, published 2021-11-17. Notable for: historical significance,sota improvement.

ViT-G/14 (LiT) from Google Research, vision model published 2021-11-15. Notable for: sota improvement.

Masked Autoencoders ViT-H from Facebook AI Research, vision model published 2021-11-11. Notable for: highly cited,sota improvement.

Projected GAN from Heidelberg University, image generation model published 2021-11-01. Notable for: sota improvement.

CodeT5-base from Salesforce, Nanyang Technological University, language model published 2021-11-01. Notable for: sota improvement.

S4 from Stanford University, language model published 2021-10-31. Notable for: sota improvement.

EfficientZero from Tsinghua University, University of California (UC) Berkeley, Shanghai Qi Zhi institute, games model published 2021-10-30. Notable for: sota improvement.

Eve from Harvard Medical School, University of Oxford, biology model published 2021-10-27. Notable for: sota improvement.

base LM+GNN+kNN from Shannon.AI, Nanjing University, Nanyang Technological University, Zhejiang University (ZJU), language model published 2021-10-17. Notable for: sota improvement.

Yuan 1.0 from Inspur, language model published 2021-10-12. Notable for: sota improvement.

Megatron-Turing NLG 530B from Microsoft, Nvidia, language model published 2021-10-11. Notable for: sota improvement,training cost.

AlphaFold-Multimer from Google DeepMind, DeepMind, biology model published 2021-10-04. Notable for: highly cited,sota improvement.

Turing ULRv5 from Microsoft, language model published 2021-09-28. Notable for: sota improvement.

TrOCR from Beihang University, Microsoft Research Asia, vision model published 2021-09-21. Notable for: sota improvement.

PLATO-XL from Baidu, language model published 2021-09-20. Notable for: sota improvement.

HyperCLOVA 204B from NAVER, language model published 2021-09-10. Notable for: sota improvement.

PermuteFormer from Peking University, language model published 2021-09-06. Notable for: sota improvement.

MEB from Microsoft, search,language model published 2021-09-04. Notable for: significant use.

FLAN 137B from Google Research, language model published 2021-09-03. Notable for: highly cited,sota improvement.

XLMR-XXL from Facebook AI Research, language model published 2021-08-17. Notable for: sota improvement.

DNABERT from Northeastern University, biology model published 2021-08-15. Notable for: sota improvement.

Zidong Taichu from Chinese Academy of Sciences, Wuhan AI Computing Center, multimodal,speech,vision,language model published 2021-08-11. Notable for: historical significance.

Jurassic-1-Jumbo from AI21 Labs, language model published 2021-08-11. Notable for: training cost.

W2v-BERT from Google Brain, Massachusetts Institute of Technology (MIT), speech model published 2021-08-07. Notable for: sota improvement.

YOLOX-X from Megvii Inc, vision model published 2021-08-06. Notable for: highly cited,sota improvement.

6-Act Tether from Facebook AI Research, Georgia Institute of Technology, robotics model published 2021-08-03. Notable for: sota improvement.

SEER from Facebook AI Research, INRIA, vision model published 2021-07-29. Notable for: sota improvement.

HuBERT from Facebook AI Research, speech model published 2021-07-27. Notable for: highly cited,sota improvement.

GOAT from DeepMind, games model published 2021-07-27. Notable for: sota improvement.

GPT-3 fine-tuned on GitHub code. Powered the first GitHub Copilot.

ERNIE 3.0 from Baidu, language model published 2021-07-05. Notable for: sota improvement.

Adaptive Input Transformer + RD from Microsoft Research Asia, Soochow University, language model published 2021-06-28. Notable for: sota improvement.

Fold2Seq from IBM, Texas A&M, biology model published 2021-06-24. Notable for: sota improvement.

EfficientNetV2-XL from Google, Google Brain, vision model published 2021-06-23. Notable for: highly cited,sota improvement.

StyleGAN3-T from Nvidia, Aalto University, image generation model published 2021-06-21. Notable for: historical significance,highly cited.

StyleGAN3-R from Nvidia, Aalto University, image generation model published 2021-06-21. Notable for: historical significance,highly cited.

Denoising Diffusion Probabilistic Models (LSUN Bedroom) from University of California (UC) Berkeley, vision model published 2021-06-11. Notable for: highly cited,sota improvement.

ALIGN from Google Research, multimodal,vision,language model published 2021-06-11. Notable for: highly cited,sota improvement.

DeBERTa from Microsoft, language model published 2021-06-10. Notable for: highly cited,sota improvement.

EMDR from Mila - Quebec AI (originally Montreal Institute for Learning Algorithms), McGill University, DeepMind, language model published 2021-06-09. Notable for: sota improvement.

CoAtNet from Google, Google Research, Google Brain, vision model published 2021-06-09. Notable for: sota improvement.

ViT-G/14 from Google Brain, Google Research, vision model published 2021-06-08. Notable for: sota improvement.

ByT5-XXL from Google, Google Research, language model published 2021-05-28. Notable for: sota improvement.

Transformer local-attention (NesT-B) from Google Cloud, Google Research, vision model published 2021-05-26. Notable for: highly cited.

CogView from Tsinghua University, Alibaba DAMO Academy, image generation model published 2021-05-26. Notable for: sota improvement.

MedBERT from Peng Cheng Laboratory, University of Texas at Houston, medicine model published 2021-05-20. Notable for: sota improvement.

ADM from OpenAI, image generation model published 2021-05-11. Notable for: highly cited,sota improvement.

ProtT5-XL-U50 from Technical University of Munich, Med AI Technology, Nvidia, Oak Ridge National Laboratory, Google, Seoul National University, biology model published 2021-05-04. Notable for: highly cited,sota improvement.

ProtBERT-BFD from Technical University of Munich, Nvidia, Seoul National University, Google, Oak Ridge National Laboratory, Med AI Technology, biology model published 2021-05-04. Notable for: sota improvement.

PLUG from Alibaba, language model published 2021-04-19. Notable for: sota improvement.

Unicorn from Allen Institute for AI, language model published 2021-03-24. Notable for: sota improvement.

M6-T from Alibaba, multimodal,language,vision model published 2021-03-05. Notable for: sota improvement.

Generative BST from Facebook AI Research, language model published 2021-03-05. Notable for: sota improvement.

Meta Pseudo Labels from Google Brain, Google AI, vision model published 2021-03-01. Notable for: sota improvement.

SRU++ Large from ASAPP, language model published 2021-02-24. Notable for: sota improvement.

Rational DQN Average from TU Darmstadt, games model published 2021-02-18. Notable for: sota improvement.

MSA Transformer from Facebook AI Research, University of California (UC) Berkeley, New York University (NYU), biology model published 2021-02-13. Notable for: sota improvement.

top-down frozen classifier from University of Edinburgh, Toshiba Cambridge Research Laboratory, speech model published 2021-02-09. Notable for: sota improvement.

DLWP from University of Washington, Microsoft Research, earth science model published 2021-02-09. Notable for: sota improvement.

DeiT-B from Meta AI, Sorbonne University, vision model published 2021-01-15. Notable for: highly cited.

Switch from Google, language model published 2021-01-11. Notable for: highly cited,sota improvement.

BigSSL from Google, Apple, speech model published 2021-01-10. Notable for: sota improvement.

12 billion parameter GPT-3 variant that draws from text. The avocado armchair.

CLIP (ResNet-50) from OpenAI, multimodal,vision,language,video model published 2021-01-05. Notable for: highly cited,sota improvement.

Image and text encoders trained to agree on 400 million pairs. Zero-shot classification and the guide inside image generators.

ERNIE-Doc (247M) from Baidu, language model published 2020-12-31. Notable for: sota improvement.

CT-MoS (WT2) from Google, National Tsing Hua University, language model published 2020-12-25. Notable for: sota improvement.

DensePhrases from Korea University, Princeton University, language model published 2020-12-23. Notable for: sota improvement.

VQGAN + CLIP from Heidelberg University, image generation model published 2020-12-17. Notable for: highly cited,sota improvement.

ESM1b from Facebook AI Research, New York University (NYU), biology model published 2020-12-15. Notable for: highly cited,sota improvement.

CPM-Large from Tsinghua University, Beijing Academy of Artificial Intelligence / BAAI, language model published 2020-12-01. Notable for: sota improvement.

Protein structure from sequence with an attention-based Evoformer. 200 million structures released.

KEPLER from Tsinghua University, Mila - Quebec AI (originally Montreal Institute for Learning Algorithms), HEC, CIFAR AI Research, Princeton University, University of Montreal / Université de Montréal, language model published 2020-11-23. Notable for: sota improvement.

wave2vec 2.0 LARGE from Facebook, speech model published 2020-10-22. Notable for: highly cited,sota improvement.

ViT-Huge/14 from Google Brain, Google Research, vision model published 2020-10-22. Notable for: highly cited.

ViT-Base/32 from Google Brain, vision model published 2020-10-22. Notable for: highly cited.

German ELECTRA Large from deepset, Bayerische Staatsbibliothek Muenchen, language model published 2020-10-21. Notable for: sota improvement.

mT5-XXL from Google, Google Research, language model published 2020-10-20. Notable for: highly cited,sota improvement.

Conformer + Wav2vec 2.0 + Noisy Student from Google, Google Research, Google Brain, speech model published 2020-10-20. Notable for: sota improvement.

LUKE from University of Washington, National Institute of Informatics, language model published 2020-10-02. Notable for: sota improvement.

ProBERTa from University of Illinois Urbana-Champaign (UIUC), Reed College, biology model published 2020-09-01. Notable for: sota improvement.

ERNIE-GEN (large) from Baidu, language model published 2020-08-06. Notable for: sota improvement.

DeLighT from University of Washington, Allen Institute for AI, Facebook AI Research, language model published 2020-08-03. Notable for: sota improvement.

EfficientDet from Google Brain, vision model published 2020-07-27. Notable for: highly cited,sota improvement.

Hopfield Networks (2020) from Johannes Kepler University Linz, Institute of Advanced Research in Artificial Intelligence, University of Oslo, biology,vision,language,medicine model published 2020-07-16. Notable for: sota improvement.

SemExp from Carnegie Mellon University (CMU), Facebook AI Research, robotics model published 2020-07-02. Notable for: sota improvement.

GShard (dense) from Google, language model published 2020-06-30. Notable for: sota improvement.

175 billion parameters. Few-shot learning from the prompt. Sold through the first LLM API.

DETR from Facebook, vision model published 2020-05-26. Notable for: highly cited.

Retrieval-Augmented Generator from Facebook, New York University (NYU), University College London (UCL), language model published 2020-05-22. Notable for: highly cited,sota improvement.

Conformer from Google, speech model published 2020-05-16. Notable for: highly cited,sota improvement.

ContextNet from Google, speech model published 2020-05-07. Notable for: sota improvement.

NAS+ESS (23M) from Northeastern University (China), NiuTrans Research, Kingsoft, language model published 2020-05-06. Notable for: sota improvement.

UnifiedQA from Allen Institute for AI, University of Washington, language model published 2020-05-02. Notable for: sota improvement.

Once for All from MIT-IBM Watson AI Lab, Massachusetts Institute of Technology (MIT), IBM, vision model published 2020-04-29. Notable for: sota improvement.

Go-explore from Uber AI, OpenAI, games model published 2020-04-27. Notable for: sota improvement.

CURL from University of California (UC) Berkeley, games model published 2020-04-08. Notable for: sota improvement.

Agent57 from DeepMind, games model published 2020-03-30. Notable for: sota improvement.

MetNet from Google, earth science model published 2020-03-24. Notable for: sota improvement.

ELECTRA from Stanford University, Google, Google Brain, language model published 2020-03-23. Notable for: discretionary.

Tensor-Transformer(1core)+PN (WT103) from University of California (UC) Berkeley, language model published 2020-03-17. Notable for: sota improvement.

Routing Transformer (WT-103) from Google Research, language model published 2020-03-12. Notable for: sota improvement.

TransformerXL + spectrum control from University of California Los Angeles (UCLA), JD.com, language model published 2020-03-11. Notable for: sota improvement.

TCAN (WT2) from Nanjing University, Ant Group, language model published 2020-02-28. Notable for: sota improvement.

Feedback Transformer from LORIA, University of Lorraine, Facebook AI Research, language model published 2020-02-21. Notable for: sota improvement.

FFN SwiGLU from Google, language model published 2020-02-14. Notable for: historical significance.

Turing-NLG from Microsoft, language model published 2020-02-13. Notable for: sota improvement.

SimCLR from Google Brain, vision model published 2020-02-13. Notable for: highly cited.

ALBERT-xxlarge from Toyota Technological Institute at Chicago, Google, language model published 2020-02-09. Notable for: highly cited.

TaLK Convolution from Carleton University, language model published 2020-02-08. Notable for: sota improvement.

Perceiver IO (optical flow) from DeepMind, multimodal,language,vision model published 2020-02-08. Notable for: sota improvement.

Theseus 6/768 from University of California San Diego, Beihang University, Microsoft, language model published 2020-02-07. Notable for: sota improvement.

Meena from Google Brain, language model published 2020-01-28. Notable for: sota improvement.

ContextNet + Noisy Student from Google, speech model published 2020-01-19. Notable for: sota improvement.

AlphaFold from DeepMind, biology model published 2020-01-15. Notable for: sota improvement,highly cited.

Big Transfer (BiT-L) from Google Brain, vision model published 2019-12-24. Notable for: sota improvement.

DD-PPO from Georgia Institute of Technology, Facebook AI Research, Oregon State University, Simon Fraser University, robotics model published 2019-12-19. Notable for: sota improvement.

OpenAI Five Rerun from OpenAI, games model published 2019-12-13. Notable for: highly cited,sota improvement.

OpenAI Five from OpenAI, games model published 2019-12-13. Notable for: highly cited,sota improvement.

MMLSTM (WT-2) from Beijing University of Posts and Telecommunications, University of West London, language model published 2019-12-05. Notable for: sota improvement.

MMLSTM (PTB) from Beijing University of Posts and Telecommunications, University of West London, language model published 2019-12-05. Notable for: sota improvement.

StarGAN v2 from NAVER, Yonsei University, Swiss Federal Institute of Technology, vision,image generation model published 2019-12-04. Notable for: highly cited,sota improvement.

StyleGAN2 from Nvidia, Aalto University, image generation model published 2019-12-03. Notable for: historical significance,highly cited.

Transformer-XL DeFINE (141M) from University of Washington, Allen Institute for AI, language model published 2019-11-27. Notable for: sota improvement.

Photo-Geometric Autoencoder from University of Oxford, 3d modeling,vision model published 2019-11-25. Notable for: sota improvement.

Transformer - LibriVox + Decoding/Rescoring from Facebook, speech model published 2019-11-19. Notable for: sota improvement.

MuZero from DeepMind, games model published 2019-11-19. Notable for: highly cited,sota improvement.

MoCo from Facebook AI, vision,image generation model published 2019-11-13. Notable for: highly cited.

Noisy Student (L2) from Carnegie Mellon University (CMU), Google, vision model published 2019-11-11. Notable for: highly cited,sota improvement.

Sandwich Transformer from Allen Institute for AI, Facebook AI Research, language model published 2019-11-10. Notable for: sota improvement.

CamemBERT from Facebook, INRIA, Sorbonne University, language model published 2019-11-10. Notable for: sota improvement.

XLM-RoBERTa from Facebook AI, language model published 2019-11-05. Notable for: highly cited,sota improvement.

Base LM + kNN LM + Continuous Cache from Stanford University, Facebook AI Research, language model published 2019-11-01. Notable for: sota improvement.

AlphaStar from DeepMind, games model published 2019-10-30. Notable for: discretionary.

BART-large from Facebook AI, language model published 2019-10-29. Notable for: highly cited.

T5-3B from Google, language model published 2019-10-23. Notable for: highly cited.

Google's text-to-text Transformer, 11 billion parameters, trained on C4.

M4-50B from Google, language model published 2019-10-11. Notable for: sota improvement.

DistilBERT from Hugging Face, language model published 2019-10-02. Notable for: highly cited.

AlphaX-1 from Facebook AI Research, Brown University, vision model published 2019-10-02. Notable for: sota improvement.

ALBERT from Toyota Technological Institute at Chicago, Google Research, language model published 2019-09-26. Notable for: highly cited.

Adaptive Inputs + LayerDrop from Facebook AI Research, LORIA, language model published 2019-09-25. Notable for: sota improvement.

Megatron-LM (8.3B) from Nvidia, language model published 2019-09-17. Notable for: highly cited,sota improvement.

Megatron-LM (1.2B) from Nvidia, language model published 2019-09-17. Notable for: highly cited.

Megatron-BERT from Nvidia, language model published 2019-09-17. Notable for: highly cited,sota improvement.

UDSMProt from Fraunhofer Heinrich Hertz Institute, biology model published 2019-09-04. Notable for: sota improvement.

Mogrifier (d2, MoS2, MC) + dynamic eval from DeepMind, University of Oxford, language model published 2019-09-04. Notable for: sota improvement.

trRosetta from Nankai University, University of Washington, Tianjin University, Harvard University, biology model published 2019-08-22. Notable for: sota improvement.

EN^2AS with performance reward from Beijing Institute of Technology, University of Technology Sydney, Monash University, language model published 2019-07-22. Notable for: sota improvement.

Pluribus from Facebook AI Research, games model published 2019-07-11. Notable for: sota improvement.

BigBiGAN from Google, vision,image generation model published 2019-07-04. Notable for: sota improvement.

RoBERTa Large from Facebook, University of Washington, language model published 2019-07-01. Notable for: highly cited,sota improvement.

Walking Minotaur robot from University of California (UC) Berkeley, Google Brain, robotics model published 2019-06-19. Notable for: sota improvement.

LaNet-L (CIFAR-10) from Brown University, Facebook, vision model published 2019-06-17. Notable for: sota improvement.

PG-SWGAN from ETH Zurich, image generation model published 2019-06-15. Notable for: sota improvement.

FixRes ResNeXt-101 WSL from Facebook AI, vision model published 2019-06-14. Notable for: sota improvement.

Char-CNN-BiLSTM from Capital One, language model published 2019-06-13. Notable for: sota improvement.

AWD-LSTM + MoS + Partial Shuffled from University of Texas at Austin, language model published 2019-06-10. Notable for: sota improvement.

Transformer-XL Large + Phrase Induction from Massachusetts Institute of Technology (MIT), University of Illinois Urbana-Champaign (UIUC), language model published 2019-06-04. Notable for: sota improvement.

XLNet from Carnegie Mellon University (CMU), Google Brain, language model published 2019-06-01. Notable for: highly cited.

XLM from Facebook, language model published 2019-06-01. Notable for: highly cited,sota improvement.

DLRM-2020 from Facebook AI, recommendation model published 2019-05-31. Notable for: sota improvement.

EfficientNet-L2 from Google, vision model published 2019-05-28. Notable for: highly cited.

CPC v2 from DeepMind, University of California (UC) Berkeley, vision model published 2019-05-22. Notable for: sota improvement.

AWD-LSTM-DRILL + dynamic evaluation† (WT2) from IDIAP, language model published 2019-05-14. Notable for: sota improvement.

ResNeXt-101 Billion-scale from Facebook AI, vision model published 2019-05-02. Notable for: sota improvement.

RaptorX-Contact from Toyota Technological Institute at Chicago, biology model published 2019-05-02. Notable for: sota improvement.

Neuro-Symbolic Concept Learner from Massachusetts Institute of Technology (MIT), Tsinghua University, MIT-IBM Watson AI Lab, DeepMind, vision,language model published 2019-04-26. Notable for: sota improvement.

MuseNet from OpenAI, audio model published 2019-04-25. Notable for: significant use.

BERT-Large-CAS (PTB+WT2+WT103) from Amazon, language model published 2019-04-20. Notable for: sota improvement.

Transformer-XL + RMS dynamic eval from University of Edinburgh, language model published 2019-04-17. Notable for: sota improvement.

WeNet (Penn Treebank) from Amazon, language model published 2019-04-08. Notable for: sota improvement.

True-Regularization+Finetune+Dynamic-Eval from Mobvoi, Williams College, language model published 2019-04-08. Notable for: sota improvement.

Cross-lingual alignment from Tel Aviv University, Massachusetts Institute of Technology (MIT), language model published 2019-04-04. Notable for: sota improvement.

SciBERT from Allen Institute for AI, language model published 2019-03-26. Notable for: highly cited,sota improvement.

NMT Transformer 437M from Google, Bar-Ilan University, language model published 2019-02-28. Notable for: sota improvement.

KataGo from Jane Street, games model published 2019-02-27. Notable for: sota improvement.

1.5 billion parameters trained on WebText. Wrote a plausible news story about unicorns and got held back for nine months.

SDE from Carnegie Mellon University (CMU), Google Brain, Monash University, language model published 2019-02-09. Notable for: sota improvement.

Hanabi 4 player from DeepMind, University of Oxford, Carnegie Mellon University (CMU), Google Brain, games model published 2019-02-01. Notable for: historical significance.

MT-DNN from Microsoft, language model published 2019-01-31. Notable for: highly cited,sota improvement.

Transformer-XL (257M) from Carnegie Mellon University (CMU), Google Brain, language model published 2019-01-09. Notable for: discretionary.

Transformer ELMo from Allen Institute for AI, University of Washington, language model published 2019-01-01. Notable for: sota improvement.

StyleGAN from Nvidia, image generation model published 2018-12-12. Notable for: historical significance,highly cited.

SPN (ImageNet 128) from Google Brain, DeepMind, image generation model published 2018-12-04. Notable for: sota improvement.

GPipe (Transformer) from Google, language model published 2018-11-16. Notable for: highly cited,sota improvement.

Multi-cell LSTM from University of Hyderabad, language model published 2018-11-15. Notable for: sota improvement.

Fine-tuned-AWD-LSTM-DOC (fin) from Samsung R&D Institute Russia, language model published 2018-11-12. Notable for: sota improvement.

Mesh-TensorFlow Transformer 4.9B (language) from Google Brain, language model published 2018-11-05. Notable for: sota improvement.

Mesh-TensorFlow Transformer 2.9B (translation) from Google Brain, language model published 2018-11-05. Notable for: sota improvement.

MemoReader from Samsung, Korea University, language model published 2018-10-31. Notable for: sota improvement.

TrellisNet from Carnegie Mellon University (CMU), Bosch Center for Artificial Intelligence, Intel Labs, language model published 2018-10-15. Notable for: sota improvement.

MetaMimic from Google, robotics model published 2018-10-11. Notable for: sota improvement.

340 million parameters, bidirectional, masked language modelling. The encoder model that ran Google Search from 2019.

Transformer (Adaptive Input Embeddings) WT103 from Facebook AI Research, language model published 2018-09-28. Notable for: sota improvement.

LSTM+NeuralCache from KU Leuven, ESAT - PSI, Apple, language model published 2018-09-24. Notable for: sota improvement.

AWD-LSTM-MoS + dynamic evaluation (WT2, 2018) from Peking University, Microsoft Research Asia, language model published 2018-09-18. Notable for: sota improvement.

Transformer + Simple Recurrent Unit from ASAPP, Cornell University, Google, Princeton University, language model published 2018-09-17. Notable for: sota improvement.

NetSurfP-2.0 from Technical University of Denmark, University of Copenhagen, Universidad Nacional de San Martín, AIMST University, biology model published 2018-09-10. Notable for: sota improvement.

(ensemble): AWD-LSTM-DOC (fin) × 5 (WT2) from NTT Communication Science Laboratories, Tohoku University, language model published 2018-08-30. Notable for: sota improvement.

Big Transformer for Back-Translation from Facebook AI Research, Google Brain, language model published 2018-08-28. Notable for: highly cited,sota improvement.

AWD-LSTM-MoS+PDR + dynamic evaluation (WT2) from IBM, language model published 2018-08-14. Notable for: sota improvement.

Big-Little Net (speech) from IBM, speech model published 2018-07-10. Notable for: sota improvement.

Big-Little Net from IBM, vision model published 2018-07-10. Notable for: sota improvement.

FTW (For The Win) from DeepMind, games model published 2018-07-03. Notable for: sota improvement.

MobileNetV2 from Google, vision model published 2018-06-18. Notable for: highly cited.

Relational Memory Core from DeepMind, University College London (UCL), language model published 2018-06-05. Notable for: sota improvement.

117 million parameters, 12 Transformer decoder layers, pretrained on BooksCorpus.

aLSTM(depth-2)+RecurrentPolicy (WT2) from University of Manchester, Alan Turing Institute, language model published 2018-05-22. Notable for: sota improvement.

Dropout-LSTM+Noise(Bernoulli) (WT2) from Columbia University, New York University (NYU), Princeton University, language model published 2018-05-03. Notable for: sota improvement.

ResNeXt-101 32x48d from Facebook, vision model published 2018-05-02. Notable for: highly cited,sota improvement.

YOLOv3 from University of Washington, vision model published 2018-04-08. Notable for: highly cited.

4 layer QRNN (h=2500) from Salesforce Research, language model published 2018-03-22. Notable for: sota improvement.

Chinese - English translation from Microsoft, language model published 2018-03-01. Notable for: sota improvement.

TCN (P-MNIST) from Carnegie Mellon University (CMU), Intel Labs, vision model published 2018-02-15. Notable for: sota improvement.

DeepLabV3+ from Google, vision model published 2018-02-07. Notable for: highly cited.

IMPALA from DeepMind, games model published 2018-02-05. Notable for: highly cited,sota improvement.

QRNN from Salesforce Research, language model published 2018-02-01. Notable for: sota improvement.

ELMo from University of Washington, Allen Institute for AI, language model published 2018-02-01. Notable for: highly cited.

T-DMCA from Google Brain, language model published 2018-01-30. Notable for: historical significance.

One algorithm for Go, chess and shogi, from self-play in hours.

DL scaling speech from Baidu, speech model published 2017-12-01. Notable for: training cost.

DL scaling LM from Baidu, language model published 2017-12-01. Notable for: historical significance.

DL scaling Image from Baidu, vision model published 2017-12-01. Notable for: training cost.

TriNet from Visual Computing Institute, RWTH Aachen University, video model published 2017-11-21. Notable for: highly cited,historical significance.

AWD-LSTM-MoS + dynamic evaluation (WT2, 2017) from Carnegie Mellon University (CMU), language model published 2017-11-10. Notable for: sota improvement.

VQ-VAE from DeepMind, vision model published 2017-11-02. Notable for: highly cited.

Fraternal dropout + AWD-LSTM 3-layer (WT2) from Jagiellonian University, Mila - Quebec AI (originally Montreal Institute for Learning Algorithms), University of Montreal / Université de Montréal, language model published 2017-10-31. Notable for: sota improvement.

DCN+ from Salesforce Research, language model published 2017-10-31. Notable for: sota improvement.

S-Norm from University of Washington, Allen Institute for AI, language model published 2017-10-29. Notable for: sota improvement.

PhraseCond from Carnegie Mellon University (CMU), University of Pittsburgh, language model published 2017-10-28. Notable for: sota improvement.

ProgressiveGAN from Nvidia, vision model published 2017-10-27. Notable for: highly cited.

AlphaGo Master from DeepMind, games model published 2017-10-19. Notable for: highly cited.

Learned Go from self-play alone, no human games, and beat AlphaGo Lee 100 to 0.

Rainbow DQN from DeepMind, games model published 2017-10-06. Notable for: sota improvement.

AWD-LSTM+WT+Cache+IOG (WT2) from NTT Communication Science Laboratories, language model published 2017-09-26. Notable for: sota improvement.

LSTM + dynamic eval from University of Edinburgh, language model published 2017-09-21. Notable for: sota improvement.

ISS from Duke University, Microsoft, language model published 2017-09-15. Notable for: sota improvement.

PyramidNet from Korea Advanced Institute of Science and Technology (KAIST), vision model published 2017-09-06. Notable for: sota improvement.

SENet (ImageNet) from Chinese Academy of Sciences, University of Oxford, vision model published 2017-09-05. Notable for: highly cited.

GL-LWGC-AWD-MoS-LSTM + dynamic evaluation (WT2) from Ben-Gurion University of the Negev, language model published 2017-08-29. Notable for: sota improvement.

Libratus from Carnegie Mellon University (CMU), games model published 2017-08-19. Notable for: sota improvement.

NeuMF (Pinterest) from Shandong University, Texas A&M, National University of Singapore, Columbia University, recommendation model published 2017-08-16. Notable for: highly cited.

EI-REHN-1000D from Korea Advanced Institute of Science and Technology (KAIST), language model published 2017-08-14. Notable for: sota improvement.

OpenAI TI7 DOTA 1v1 from OpenAI, games model published 2017-08-11. Notable for: historical significance,sota improvement.

RetinaNet-R50 from Facebook AI Research, vision model published 2017-08-07. Notable for: highly cited.

RetinaNet-R101 from Facebook AI Research, vision model published 2017-08-07. Notable for: highly cited.

AWD-LSTM - 3-layer LSTM (tied) + continuous cache pointer (WT2) from Salesforce Research, language model published 2017-08-07. Notable for: highly cited,sota improvement.

GSM from Peking University, Microsoft Research, language model published 2017-07-30. Notable for: sota improvement.

ConvS2S (ensemble of 8 models) from Meta AI, language model published 2017-07-25. Notable for: highly cited,sota improvement.

PSPNet from Chinese University of Hong Kong (CUHK), vision model published 2017-07-21. Notable for: highly cited.

NASNet-A from Google Brain, vision model published 2017-07-21. Notable for: highly cited.

AWD-LSTM from DeepMind, University of Oxford, language model published 2017-07-18. Notable for: sota improvement.

JFT from Google Research, Carnegie Mellon University (CMU), vision model published 2017-07-10. Notable for: sota improvement.

DeepLoc from Technical University of Denmark, University of Copenhagen, biology model published 2017-07-07. Notable for: sota improvement.

ShuffleNet v1 from Megvii Inc, vision model published 2017-07-03. Notable for: highly cited.

NoisyNet-Dueling from DeepMind, games model published 2017-06-30. Notable for: sota improvement.

DeepLabV3 from Google, vision model published 2017-06-17. Notable for: highly cited.

HRA from Maluuba, Microsoft, games model published 2017-06-13. Notable for: sota improvement.

Transformer from Google Research, Google Brain, language model published 2017-06-12. Notable for: highly cited,historical significance.

EDSR from Seoul National University, vision,image generation model published 2017-06-10. Notable for: highly cited.

Reading Twice for NLU from DeepMind, language model published 2017-06-08. Notable for: sota improvement.

PointNet++ from Stanford University, 3d modeling model published 2017-06-07. Notable for: highly cited.

Inflated 3D ConvNet from DeepMind, University of Oxford, 3d modeling model published 2017-06-01. Notable for: highly cited.

SRGAN from Twitter, vision,image generation model published 2017-05-25. Notable for: highly cited.

Mnemonic Reader from Fudan University, Microsoft Research, language model published 2017-05-08. Notable for: sota improvement.

DeepLab (2017) from Johns Hopkins University, Google, University College London (UCL), vision model published 2017-04-27. Notable for: highly cited.

MobileNet from Google, vision model published 2017-04-17. Notable for: highly cited.

WGAN-GP from Courant Institute of Mathematical Sciences, Mila - Quebec AI (originally Montreal Institute for Learning Algorithms), image generation model published 2017-03-31. Notable for: highly cited.

Mask R-CNN from Facebook AI Research, vision model published 2017-03-30. Notable for: highly cited.

Prototypical networks from University of Toronto, Twitter, vision model published 2017-03-15. Notable for: highly cited.

DnCNN from Harbin Institute of Technology, Hong Kong Polytechnic University, ULSee Inc., Xi’an Jiaotong University, vision,image generation model published 2017-02-01. Notable for: highly cited.

MoE-Multi from Jagiellonian University, Google Brain, language model published 2017-01-23. Notable for: highly cited,sota improvement,historical significance.

OR-WideResNet from Duke University, University of Chinese Academy of Sciences, vision model published 2017-01-07. Notable for: sota improvement.

DeepStack from University of Alberta, Charles University, Czech Technical University, games model published 2017-01-06. Notable for: sota improvement.

YOLOv2 from University of Washington, Allen Institute for AI, vision model published 2016-12-25. Notable for: highly cited.

EnhanceNet from Max Planck Institute for Intelligent Systems, vision model published 2016-12-23. Notable for: sota improvement,highly cited.

3DMM-CNN from University of Southern California, vision model published 2016-12-15. Notable for: sota improvement.

HR-ResNet101 from Carnegie Mellon University (CMU), vision model published 2016-12-13. Notable for: sota improvement.

GAN-Advancer from OpenAI, vision model published 2016-12-05. Notable for: highly cited.

PointNet from Stanford University, 3d modeling model published 2016-12-02. Notable for: highly cited.

Elastic weight consolidation from DeepMind, vision,games model published 2016-12-02. Notable for: highly cited.

Image-to-image cGAN from University of California (UC) Berkeley, vision,image generation model published 2016-11-21. Notable for: highly cited.

PolyNet from Chinese University of Hong Kong (CUHK), vision model published 2016-11-17. Notable for: sota improvement.

ResNeXt-50 from University of California San Diego, Facebook, vision model published 2016-11-16. Notable for: highly cited.

ResNeXt-101 (64×4d) from University of California San Diego, Facebook, vision model published 2016-11-16. Notable for: highly cited,sota improvement.

DAC-CSR from Jiangnan University, University of Surrey, vision model published 2016-11-16. Notable for: sota improvement.

DTN (Domain Transfer Network) from Facebook AI Research, vision,image generation model published 2016-11-07. Notable for: highly cited.

DLDL (PASCAL) from University of Oxford, vision model published 2016-11-06. Notable for: sota improvement.

NASv3 (CIFAR-10) from Google Brain, vision model published 2016-11-05. Notable for: highly cited.

NAS with base 8 and shared embeddings from Google Brain, language model published 2016-11-05. Notable for: highly cited.

BIDAF from University of Washington, Allen Institute for AI, language model published 2016-11-05. Notable for: highly cited,sota improvement.

VD-LSTM+REAL Large from Salesforce Research, Stanford University, language model published 2016-11-04. Notable for: sota improvement.

SPIDER2 from Griffith University, University of Iowa, Dezhou University, biology model published 2016-10-28. Notable for: sota improvement.

GAWWN from University of Michigan, Max Planck Institute for Informatics, vision model published 2016-10-08. Notable for: sota improvement.

Xception from Google, vision model published 2016-10-07. Notable for: highly cited.

Pointer Sentinel-LSTM (medium) from MetaMind Inc, Salesforce, language model published 2016-09-26. Notable for: highly cited,sota improvement.

GNMT from Google, language model published 2016-09-26. Notable for: highly cited.

Wide Residual Network from Université Paris-Est, vision model published 2016-09-19. Notable for: highly cited.

ResNet-200 from Microsoft Research Asia, vision model published 2016-09-17. Notable for: highly cited.

ResNet-1001 from Microsoft, vision model published 2016-09-17. Notable for: highly cited.

Youtube recommendation model from Google, recommendation model published 2016-09-15. Notable for: highly cited,historical significance,significant use.

WaveNet from Google DeepMind, speech model published 2016-09-12. Notable for: highly cited.

MS-ensemble-speech-recognition from Microsoft, speech model published 2016-09-12. Notable for: sota improvement.

LF-MMI from Johns Hopkins University, Cornell University, speech model published 2016-09-08. Notable for: highly cited,sota improvement.

DenseNet-264 from Tsinghua University, Facebook AI Research, Cornell University, vision model published 2016-08-25. Notable for: highly cited.

SimpleNet from Sensifai, Islamic Azad University, Technicolor R&I, Institute for Research in Fundamental Sciences (IPM), vision model published 2016-08-22. Notable for: sota improvement.

Character-enriched word2vec from Facebook AI Research, language model published 2016-07-15. Notable for: highly cited.

CCL from SenseTime, Chinese University of Hong Kong (CUHK), Chinese Academy of Sciences, vision model published 2016-06-27. Notable for: sota improvement.

R-FCN from Tsinghua University, Microsoft Research, vision model published 2016-06-21. Notable for: highly cited.

Segmental RNN from University of Edinburgh, Carnegie Mellon University (CMU), University of Washington, speech model published 2016-06-20. Notable for: sota improvement.

CMS-RCNN from IEEE, vision model published 2016-06-17. Notable for: sota improvement.

PixelCNN from Google DeepMind, vision model published 2016-06-16. Notable for: highly cited,sota improvement.

LRR-4X from UC Irvine, vision model published 2016-05-08. Notable for: sota improvement.

Gated HORNN (3rd order) from York University, language model published 2016-04-30. Notable for: sota improvement.

Dueling DQN from Google DeepMind, games model published 2016-04-05. Notable for: sota improvement.

Template Adaptation from University of Oxford, vision model published 2016-03-12. Notable for: sota improvement.

Named Entity Recognition model from Carnegie Mellon University (CMU), language model published 2016-03-04. Notable for: discretionary.

Double DQN from Google DeepMind, games model published 2016-03-02. Notable for: sota improvement.

SqueezeNet from DeepScale, University of California (UC) Berkeley, Stanford University, vision model published 2016-02-24. Notable for: highly cited.

Inceptionv4 from Google, vision model published 2016-02-23. Notable for: highly cited.

Inception-ResNet-V2 from Google, vision model published 2016-02-23. Notable for: highly cited.

A3C FF hs from Google, University of Montreal / Université de Montréal, games model published 2016-02-04. Notable for: highly cited,sota improvement.

The version that beat Lee Sedol in March 2016. Policy and value networks plus tree search, trained on human games then self-play.

Variational (untied weights, MC) LSTM (Large) from University of Cambridge, language model published 2015-12-16. Notable for: highly cited,sota improvement.

Advantage Learning from Google DeepMind, games model published 2015-12-15. Notable for: sota improvement.

The 152-layer residual network that won ImageNet 2015 with 3.6% top-5 error, below the human level.

ResNet-110 (CIFAR-10) from Microsoft, vision model published 2015-12-10. Notable for: highly cited.

ResNet-101 (ImageNet) from Microsoft, vision model published 2015-12-10. Notable for: highly cited.

SSD, vision model published 2015-12-08. Notable for: highly cited.

Inception v3 from Google, University College London (UCL), vision model published 2015-12-02. Notable for: highly cited.

Netflix Recommender System from Netflix, recommendation model published 2015-12-01. Notable for: highly cited,significant use,historical significance.

Multi-scale Dilated CNN from Princeton University, Intel Labs, vision model published 2015-11-23. Notable for: highly cited.

3DDFA from Chinese Academy of Sciences, Michigan State University, vision model published 2015-11-23. Notable for: sota improvement.

SAF R-CNN from Beijing Institute of Technology, Sun Yat-sen University, Panasonic R&D, National University of Singapore, vision model published 2015-10-28. Notable for: sota improvement.

AlphaGo Fan from DeepMind, games model published 2015-10-01. Notable for: highly cited,sota improvement.

Deep Deterministic Policy Gradients from Google DeepMind, robotics model published 2015-09-09. Notable for: highly cited.

BPE from University of Edinburgh, language model published 2015-08-31. Notable for: highly cited.

DCNN from University of Maryland, Rutgers University, vision model published 2015-08-07. Notable for: sota improvement.

Deep CNN + COTS from IEEE, vision model published 2015-07-26. Notable for: sota improvement.

CompACT-Deep from University of California San Diego, vision model published 2015-07-19. Notable for: sota improvement.

BatchNorm from Google, vision model published 2015-06-15. Notable for: highly cited.

YOLO from University of Washington, Allen Institute for AI, Facebook AI Research, vision model published 2015-06-08. Notable for: highly cited.

CFSS from SenseTime, Chinese University of Hong Kong (CUHK), Shenzhen Institute of Advanced Technology, vision model published 2015-06-07. Notable for: sota improvement.

Faster R-CNN from Microsoft Research, vision model published 2015-06-04. Notable for: highly cited.

U-Net from University of Freiburg, vision model published 2015-05-18. Notable for: sota improvement,highly cited.

Fast R-CNN from Microsoft Research, vision model published 2015-04-30. Notable for: highly cited.

TC-DNN-BLSTM-DNN from Carnegie Mellon University (CMU), speech model published 2015-04-06. Notable for: sota improvement.

genCNN + dyn eval from Chinese Academy of Sciences, Huawei Noah's Ark Lab, Dublin City University, language model published 2015-03-17. Notable for: sota improvement.

DQN-2015 from Google, games model published 2015-02-25. Notable for: highly cited.

TRPO from University of California (UC) Berkeley, games model published 2015-02-19. Notable for: highly cited.

MSRA (C, PReLU) from Microsoft Research, vision model published 2015-02-06. Notable for: highly cited.

VGG-Face from University of Oxford, vision model published 2015-01-01. Notable for: sota improvement,highly cited.

ADAM (CIFAR-10) from University of Amsterdam, OpenAI, University of Toronto, vision model published 2014-12-22. Notable for: highly cited.

Fractional Max-Pooling from University of Warwick, vision model published 2014-12-18. Notable for: sota improvement.

SNM-skip from Google, language model published 2014-12-03. Notable for: sota improvement.

TA-CNN from Chinese University of Hong Kong (CUHK), vision model published 2014-11-29. Notable for: sota improvement.

Cascaded LNet-ANet from Chinese University of Hong Kong (CUHK), vision model published 2014-11-28. Notable for: highly cited.

Fully Convolutional Networks from University of California (UC) Berkeley, vision model published 2014-11-14. Notable for: highly cited.

SC-NLM from University of Toronto, multimodal,vision,language model published 2014-11-10. Notable for: highly cited.

LRCN from UT Austin, University of Massachusetts Lowell, University of California (UC) Berkeley, video model published 2014-11-07. Notable for: highly cited.

Spatially-Sparse CNN from University of Warwick, vision model published 2014-09-23. Notable for: sota improvement.

Deeply-supervised nets from Microsoft Research, vision model published 2014-09-18. Notable for: highly cited,sota improvement.

GoogLeNet / InceptionV1 from Google, University of Michigan, University of North Carolina, vision model published 2014-09-17. Notable for: highly cited.

SPN-4+KN5 from Singapore University of Technology & Design, DSO National Laboratories, language model published 2014-09-14. Notable for: sota improvement.

Seq2Seq LSTM from Google, language model published 2014-09-10. Notable for: highly cited.

VGG19 from University of Oxford, vision model published 2014-09-04. Notable for: highly cited.

VGG16 from University of Oxford, vision model published 2014-09-04. Notable for: highly cited.

RNNsearch-50* from Jacobs University Bremen, University of Montreal / Université de Montréal, language model published 2014-09-01. Notable for: highly cited.

NPD from IEEE, vision model published 2014-08-06. Notable for: sota improvement.

ACF-WIDER from Chinese Academy of Sciences, vision model published 2014-07-15. Notable for: sota improvement.

SmooCT from University College London (UCL), games model published 2014-07-01. Notable for: sota improvement.

Multiresolution CNN from Google, Stanford University, video model published 2014-06-23. Notable for: highly cited.

DeepFace from Tel Aviv University, Facebook, vision model published 2014-06-23. Notable for: highly cited.

RNN-WER from DeepMind, University of Toronto, speech model published 2014-06-22. Notable for: highly cited,sota improvement.

Fragment embedding from Stanford University, vision model published 2014-06-21. Notable for: sota improvement.

SPPNet from Microsoft, Xi’an Jiaotong University, University of Science and Technology of China (USTC), vision model published 2014-06-18. Notable for: highly cited.

GANs from University of Montreal / Université de Montréal, image generation model published 2014-06-10. Notable for: highly cited.

Two-stream ConvNets for action recognition from University of Oxford, video model published 2014-06-09. Notable for: highly cited.

GRUs from University of Montreal / Université de Montréal, Jacobs University, University of Maine, language model published 2014-06-03. Notable for: highly cited.

Dropout: SVHN from University of Toronto, vision model published 2014-06-01. Notable for: highly cited,sota improvement.

AdaRNN from Beihang University, language model published 2014-06-01. Notable for: highly cited.

Paragraph Vector from Google, language model published 2014-05-14. Notable for: highly cited.

HyperNEAT from University of Texas at Austin, games model published 2014-03-05. Notable for: sota improvement.

GloVe (6B) from Stanford University, language model published 2014-01-01. Notable for: highly cited.

GloVe (32B) from Stanford University, language model published 2014-01-01. Notable for: highly cited.

OverFeat from New York University (NYU), vision model published 2013-12-21. Notable for: highly cited.

Image generation from University of Amsterdam, vision model published 2013-12-20. Notable for: highly cited.

DeepMind's Atari agent. A CNN plus Q-learning plus experience replay.

Network in Network from National University of Singapore, vision model published 2013-12-16. Notable for: highly cited.

RNN for 1B words from Google, language model published 2013-12-11. Notable for: highly cited,sota improvement.

TransE from Universite de Technologie de Compiègne – CNRS, Google, language model published 2013-12-05. Notable for: highly cited.

DeViSE from Google, vision model published 2013-12-05. Notable for: highly cited,sota improvement.

Visualizing CNNs from New York University (NYU), vision model published 2013-11-12. Notable for: highly cited.

R-CNN (T-net) from University of California (UC) Berkeley, vision model published 2013-11-11. Notable for: highly cited.

Word2Vec (small) from Google, language model published 2013-10-16. Notable for: highly cited.

Google's word embeddings, trained on 100 billion words of news.

RNTN from Stanford University, language model published 2013-10-01. Notable for: highly cited.

RCTM from University of Oxford, language model published 2013-10-01. Notable for: highly cited.

Hierarchical Scene Labeling (Stanford Background) from New York University (NYU), vision model published 2013-08-01. Notable for: highly cited,historical significance.

ReLU-Speech from Google, University of Toronto, New York University (NYU), speech model published 2013-05-26. Notable for: training cost,sota improvement.

Multilingual DNN from Google, speech model published 2013-05-26. Notable for: sota improvement,training cost.

Selective Search from University of Trento, University of Amsterdam, vision model published 2013-04-02. Notable for: highly cited.

PreTrans-3L-250H from University of Toronto, speech model published 2013-03-22. Notable for: highly cited.

DistBelief NNLM from Google, language model published 2013-01-16. Notable for: highly cited,sota improvement.

DNN EM segmentation from IDSIA, SUPSI, vision model published 2012-12-03. Notable for: historical significance,highly cited.

DistBelief Vision from Google, vision model published 2012-12-03. Notable for: sota improvement,highly cited.

DistBelief Speech from Google, speech model published 2012-12-03. Notable for: highly cited.

Bayesian automated hyperparameter tuning from University of Toronto, University of Sherbrooke, Harvard University, other model published 2012-12-02. Notable for: highly cited.

RNN+LDA+KN5+cache from Microsoft, Brno University of Technology, language model published 2012-12-01. Notable for: sota improvement.

Eight layers, 60 million parameters, two GTX 580s for six days. Won ImageNet 2012 and started the deep learning era.

LSTM LM from RWTH Aachen University, language model published 2012-09-09. Notable for: highly cited.

Context-dependent RNN from Microsoft Research, Brno University of Technology, language model published 2012-07-27. Notable for: sota improvement.

Unsupervised High-level Feature Learner from Google, vision model published 2012-07-12. Notable for: highly cited,sota improvement.

Dropout (TIMIT) from University of Toronto, speech model published 2012-06-03. Notable for: highly cited.

Dropout (MNIST) from University of Toronto, vision model published 2012-06-03. Notable for: highly cited.

Dropout (ImageNet) from University of Toronto, vision model published 2012-06-03. Notable for: highly cited.

Dropout (CIFAR) from University of Toronto, vision model published 2012-06-03. Notable for: highly cited.

NLP from scratch from NEC Laboratories, Princeton University, language model published 2011-11-08. Notable for: highly cited.

CNN committee (traffic sign) from IDSIA, vision model published 2011-10-03. Notable for: historical significance,sota improvement.

Adaptive Subgrad from Technion - Israel Institute of Technology, Google, University of California (UC) Berkeley, language model published 2011-10-03. Notable for: highly cited.

CNN Committee (NIST) from IDSIA, vision model published 2011-09-18. Notable for: historical significance.

CNN Committee (MNIST) from IDSIA, vision model published 2011-09-18. Notable for: historical significance.

High Performance CNN (NORB) from IDSIA, SUPSI, vision model published 2011-07-16. Notable for: historical significance,sota improvement.

Recursive Neural Network from Stanford University, vision,language model published 2011-06-28. Notable for: highly cited.

Vector Space Model from Stanford University, language model published 2011-06-19. Notable for: highly cited,sota improvement.

Cross-Lingual POS Tagger from Carnegie Mellon University (CMU), Google Research, language model published 2011-06-19. Notable for: sota improvement.

Deep Autoencoders from University of Toronto, vision model published 2011-04-29. Notable for: historical significance.

Deep rectifier networks from University of Montreal / Université de Montréal, vision model published 2011-04-13. Notable for: highly cited.

YouTube Video Recommendation System from Google, recommendation model published 2010-09-26. Notable for: highly cited,significant use,historical significance.

RNN LM from Johns Hopkins University, language model published 2010-09-26. Notable for: highly cited.

Pooling CNN (NORB) from University of Bonn, vision model published 2010-09-15. Notable for: historical significance,highly cited,sota improvement.

Pooling CNN (Caltech 101) from University of Bonn, vision model published 2010-09-15. Notable for: historical significance,highly cited,sota improvement.

ReLU (NORB) from University of Toronto, vision model published 2010-06-15. Notable for: highly cited.

ReLU (LFW) from University of Toronto, vision model published 2010-06-15. Notable for: highly cited.

iCCCP from Massachusetts Institute of Technology (MIT), vision model published 2010-06-13. Notable for: historical significance,sota improvement.

Feedforward NN from University of Montreal / Université de Montréal, vision model published 2010-05-13. Notable for: highly cited.

Stacked Denoising Autoencoders from University of Montreal / Université de Montréal, University of Toronto, other model published 2010-01-03. Notable for: highly cited.

Super-vector coding from University of Illinois Urbana-Champaign (UIUC), NEC Laboratories, Rutgers University, vision model published 2010-01-01. Notable for: sota improvement.

LCNP NORB, vision model published 2009-11-22. Notable for: historical significance.

LCNP MNIST, vision model published 2009-11-22. Notable for: historical significance.

LCNP LabelMe from University of Bonn, vision model published 2009-11-22. Notable for: historical significance.

BellKor 2007 from AT&T, recommendation model published 2009-09-21. Notable for: sota improvement.

Two Stage Feature Extraction (MNIST) from New York University (NYU), vision model published 2009-09-01. Notable for: highly cited,sota improvement.

MatrixFac for Recommenders from Yahoo Research, AT&T, recommendation model published 2009-08-07. Notable for: highly cited.

Pragmatic Theory solution (Netflix 2009) from Pragmatic Theory Inc., recommendation model published 2009-08-01. Notable for: sota improvement.

BigChaos OptiBlend from AT&T, recommendation model published 2009-08-01. Notable for: sota improvement.

BellKor 2009 from AT&T, recommendation model published 2009-08-01. Notable for: historical significance.

BellKor 2008 from AT&T, recommendation model published 2009-08-01. Notable for: sota improvement.

GPU DBNs from Stanford University, other model published 2009-06-15. Notable for: highly cited,historical significance.

RBM Image Classifier from University of Toronto, vision model published 2009-04-08. Notable for: highly cited.

BP-DBN from University of Toronto, speech model published 2009-01-01. Notable for: historical significance,sota improvement.

GNN from University of Siena, other model published 2008-12-09. Notable for: historical significance,highly cited.

HLBL from University of Toronto, language model published 2008-12-08. Notable for: highly cited.

BigChaos 2008 from AT&T, recommendation model published 2008-11-25. Notable for: historical significance.

Sparse digit recognition SVM from University of Lubeck, vision model published 2008-11-19. Notable for: sota improvement.

Denoising Autoencoders from University of Montreal / Université de Montréal, other model published 2008-07-05. Notable for: highly cited.

Deep Multitask NLP Network from NEC Laboratories, language model published 2008-07-05. Notable for: highly cited,sota improvement.

BLSTM for handwriting (2) from University of Bern, IDSIA, Technical University of Munich, vision model published 2007-12-03. Notable for: sota improvement.

Enhanced Neighborhood-Based Filtering from AT&T, recommendation model published 2007-10-28. Notable for: sota improvement.

BLSTM for handwriting (1) from University of Bern, IDSIA, Technical University of Munich, vision model published 2007-09-23. Notable for: sota improvement.

SB-LM from Google, language model published 2007-06-22. Notable for: training cost,highly cited.

KN-LM from Google, language model published 2007-06-22. Notable for: training cost,highly cited.

λ-WASP from UT Austin, language model published 2007-06-01. Notable for: sota improvement.

Greedy layer-wise DNN training from University of Montreal / Université de Montréal, other model published 2006-12-04. Notable for: highly cited.

Local Binary Patterns for facial recognition from University of Oulu, IEEE, vision model published 2006-12-01. Notable for: highly cited.

Dimensionality Reduction from University of Toronto, vision model published 2006-07-18. Notable for: highly cited.

Deep Belief Nets from University of Toronto, National University of Singapore, vision model published 2006-07-18. Notable for: highly cited.

SVM-CNN from New York University (NYU), vision model published 2006-06-17. Notable for: historical significance.

Spatial Pyramid Matching from INRIA, University of Illinois Urbana-Champaign (UIUC), Ecole Normale Supèrieure, vision model published 2006-06-17. Notable for: highly cited.

Crazy Stone from INRIA, games model published 2006-05-29. Notable for: sota improvement.

FAST from University of Cambridge, video model published 2006-05-07. Notable for: highly cited.

TFE SVM from Centre de Recherche en Automatique de Nancy (CRAN), CENPARMI, vision model published 2006-02-02. Notable for: sota improvement.

RankNet from Microsoft Research, Microsoft, search model published 2005-08-07. Notable for: highly cited.

Histograms of Oriented Gradients from INRIA, vision model published 2005-06-25. Notable for: highly cited.

Hierarchical LM, language model published 2005-01-06. Notable for: highly cited.

LMICA, vision model published 2004-12-01. Notable for: training cost,historical significance.

Invariant CNN from New York University (NYU), vision model published 2004-06-27. Notable for: highly cited,historical significance.

NPLM (Brown) from University of Montreal / Université de Montréal, language model published 2003-03-15. Notable for: highly cited.

NPLM (AP News) from University of Montreal / Université de Montréal, language model published 2003-03-15. Notable for: highly cited.

LDA from Stanford University, language model published 2003-02-02. Notable for: highly cited.

Statistical Shape Constellations from California Institute of Technology, vision model published 2003-01-01. Notable for: historical significance.

Thumbs Up? from Cornell University, IBM, language model published 2002-05-28. Notable for: highly cited.

Decision tree (classification) from Mitsubishi Electric Research Labs, Compaq CRL, vision model published 2001-12-08. Notable for: highly cited.

Gradient Boosting Machine from Stanford University, mathematics model published 2001-10-01. Notable for: highly cited.

Immediate trihead from Brown University, language model published 2001-07-06. Notable for: sota improvement.

PoE MNIST from University College London (UCL), vision model published 2000-11-28. Notable for: historical significance.

Neural LM from University of Montreal / Université de Montréal, language model published 2000-11-28. Notable for: training cost,historical significance,highly cited.

IBM Model 4 from University of Southern California, IBM, University of Pennsylvania, language model published 1999-07-02. Notable for: highly cited,historical significance.

RECONTRA-uncategorized, language model published 1999-06-02. Notable for: training cost.

RECONTRA-categorized, language model published 1999-06-02. Notable for: training cost.

LSTM with forget gates from IDSIA, language model published 1999-01-02. Notable for: highly cited.

LeCun's seven-layer CNN that read cheques for banks in the late 1990s.

LSTM from Technical University of Munich, language model published 1997-11-15. Notable for: highly cited.

Bidirectional RNN from Advanced Telecommunications Research Institute, speech model published 1997-11-01. Notable for: highly cited.

IBM's chess machine. 200 million positions a second with hand-tuned evaluation.

SOM-CNN, vision model published 1997-01-31. Notable for: historical significance,highly cited.

AdaBoost.M2 Digit Recognition from AT&T, vision model published 1996-07-03. Notable for: highly cited.

System 11 from Carnegie Mellon University (CMU), vision model published 1996-06-18. Notable for: highly cited.

MUSIC perceptron, vision model published 1996-06-03. Notable for: historical significance.

LISSOM from University of Texas at Austin, vision model published 1995-11-27. Notable for: sota improvement,historical significance.

Support Vector Machines from AT&T, Bell Laboratories, vision model published 1995-09-01. Notable for: highly cited.

Predictive Coding NN from Technical University of Munich, language model published 1994-12-02. Notable for: historical significance.

NeuroChess, games model published 1994-12-02. Notable for: historical significance,highly cited.

Mixture of linear models, vision model published 1994-12-02. Notable for: historical significance.

JPMAX, vision model published 1994-12-02. Notable for: historical significance.

GroupLens from Massachusetts Institute of Technology (MIT), recommendation model published 1994-10-22. Notable for: highly cited.

Ceramic-MLP from Sapienza Università di Roma, materials science model published 1994-01-07. Notable for: historical significance.

ANN Eye Tracker, vision model published 1993-11-29. Notable for: historical significance.

Siamese-TDNN from Bell Laboratories, vision model published 1993-08-01. Notable for: historical significance,highly cited.

IBM-5 from IBM, language model published 1993-06-15. Notable for: highly cited.

Boosting from Bell Laboratories, vision model published 1992-11-30. Notable for: historical significance.

Cancer drug mechanism prediction from National Cancer Institute, medicine model published 1992-10-16. Notable for: historical significance.

Golem from Alan Turing Institute, biology model published 1992-10-01. Notable for: historical significance,sota improvement.

TD-Gammon from IBM, games model published 1992-05-01. Notable for: highly cited,historical significance.

Weight Decay, speech model published 1991-12-02. Notable for: highly cited,historical significance.

RAAM, other model published 1990-11-01. Notable for: highly cited.

SexNet compression, vision model published 1990-10-01. Notable for: historical significance.

SexNet classification, vision model published 1990-10-01. Notable for: historical significance,highly cited.

ISR network from Stanford University, vision model published 1990-10-01. Notable for: historical significance.

Bankruptcy-NN, other model published 1990-06-17. Notable for: historical significance,highly cited.

NETtalk reimplementation from Oregon State University, speech model published 1990-06-01. Notable for: historical significance,training cost.

Zip CNN from AT&T, Bell Laboratories, vision model published 1989-12-01. Notable for: highly cited.

Speaker-independent vowel classification from University of Washington, speech model published 1989-11-27. Notable for: historical significance.

Handwritten digit recognition network from AT&T, vision model published 1989-11-27. Notable for: historical significance.

Truck backer-upper from Stanford University, driving model published 1989-06-18. Notable for: historical significance.

Invariant image recognition from Complutense University of Madrid, vision model published 1989-06-18. Notable for: historical significance.

Q-learning from University of London, robotics,games model published 1989-01-01. Notable for: highly cited.

MLP baggage detector from Science Applications International Corporation / SAIC, vision model published 1989-01-01. Notable for: historical significance.

MLN-ASR from McGill University, speech model published 1988-08-01. Notable for: historical significance.

Latent semantic analysis from University of Chicago, Bell Laboratories, University of Western Ontario, language model published 1988-04-05. Notable for: historical significance.

Translation-invariant MLP from Carnegie Mellon University (CMU), published 1987-06-15. Notable for: historical significance.

NetTalk (transcription) from Princeton University, speech model published 1987-06-06. Notable for: highly cited,historical significance.

NetTalk (dictionary) from Princeton University, speech model published 1987-06-06. Notable for: highly cited,historical significance.

Optimized Multi-Scale Edge Detection from Massachusetts Institute of Technology (MIT), vision model published 1986-11-01. Notable for: highly cited.

MLP with back-propagation from University of California San Diego, Carnegie Mellon University (CMU), mathematics model published 1986-10-01. Notable for: highly cited.

Distributed representation NN from Carnegie Mellon University (CMU), other model published 1986-08-15. Notable for: historical significance,highly cited.

Error Propagation from University of California San Diego, Carnegie Mellon University (CMU), other model published 1986-01-03. Notable for: highly cited.

Hierarchical Cognitron from NHK Broadcasting Science Research Laboratories, other model published 1984-04-01. Notable for: historical significance.

ASE+ACE from University of Massachusetts Amherst, robotics model published 1983-09-01. Notable for: highly cited,historical significance.

Hopfield network from California Institute of Technology, other model published 1982-04-01. Notable for: highly cited.

Kohonen network from Helsinki University of Technology, mathematics model published 1981-07-25. Notable for: highly cited.

Neocognitron from NHK Broadcasting Science Research Laboratories, vision model published 1980-04-01. Notable for: highly cited.

Transfer Learning from University of Zagreb, vision model published 1976-07-01. Notable for: historical significance.

Cognitron from Biological Cybernetics, other model published 1975-09-01. Notable for: historical significance.

Piecewise linear model from University of Kansas, vision model published 1973-11-01. Notable for: historical significance,highly cited.

Decision tree adaline from Tokyo Medical and Dental University, medicine model published 1969-05-01. Notable for: historical significance.

GLEE from University of Edinburgh, games model published 1968-07-01. Notable for: historical significance.

Boxes (pole) from University of Edinburgh, games model published 1968-07-01. Notable for: historical significance.

LTE speaker verification system from IBM, speech model published 1966-11-01. Notable for: historical significance.

Heuristic Reinforcement Learning from Purdue University, robotics model published 1965-10-01. Notable for: historical significance,highly cited.

Print Recognition Logic from IBM, vision model published 1963-01-01. Notable for: historical significance.

MADALINE I from Stanford University, other model published 1962-07-01. Notable for: historical significance.

Linear Decision Functions from Bell Laboratories, mathematics model published 1962-06-01. Notable for: historical significance,highly cited.

ADALINE from Stanford University, vision model published 1960-06-30. Notable for: highly cited.

Perceptron (1960) from Cornell Aeronautical Laboratory, vision model published 1960-03-30. Notable for: historical significance.

Pattern recognition and reading by machine from Sandia Corporation, vision model published 1959-12-01. Notable for: historical significance.

Learned to play checkers better than its author on an IBM 701.

Pandemonium (morse) from Massachusetts Institute of Technology (MIT), language model published 1959-02-01. Notable for: highly cited,historical significance.

Rosenblatt's hardware perceptron at Cornell Aeronautical Laboratory, 400 photocells and motor-driven potentiometers as weights.

Sequence-based pattern recognition from Massachusetts Institute of Technology (MIT), vision model published 1955-03-01. Notable for: historical significance.

Self Organizing System from Massachusetts Institute of Technology (MIT), other model published 1955-03-01. Notable for: historical significance.

Genetic algorithm from Institute for Advanced Study, mathematics,biology model published 1954-07-02. Notable for: historical significance.

SNARC from Harvard University, robotics model published 1952-01-08. Notable for: historical significance.

Theseus from Bell Laboratories, robotics model published 1950-07-02. Notable for: historical significance.