Models
Models from the labs in Epoch AI's catalog, the repositories they published on Hugging Face, and posts from their own news pages.
Cosmos-Reason1 56BNvidia · Multimodal, Language, Vision, Robotics, Video · 56B parametersUnreleasedMay 19, 2025Cosmos-Reason1 7BNvidia · Multimodal, Language, Vision, Robotics, Video · 7B parametersOpen weights (restricted use)May 19, 2025NVIDIA Isaac GR00T N1.5 3BNvidia · Robotics, Vision, Language · 3B parametersOpen weights (non-commercial)May 18, 2025SANA 1.5 4.8BNvidia, Massachusetts Institute of Technology (MIT), Tsinghua University, Playground, Peking University, The University of Hong Kong · Image generation · 4.8B parametersOpen weights (non-commercial)May 17, 2025codex-1OpenAI · LanguageAPI accessMay 16, 2025codex-miniOpenAI · LanguageAPI accessMay 16, 2025AlphaEvolveDeepMind · LanguageMay 14, 2025Minimax-Speech-02-HDMiniMax · SpeechAPI accessMay 12, 2025Earth-2 (cBottle-SR)Nvidia · Earth science, Image generation · 330M parameters · Training compute 1.6e21 FLOPOpen weights (non-commercial)May 10, 2025Mistral Medium 3Mistral AI · Multimodal, Language, VisionAPI accessMay 7, 2025Apriel Nemotron 15BNvidia, ServiceNow · Language · 15B parameters · Training compute 9e21 FLOPOpen weights (unrestricted)May 6, 2025Gemini 2.5 Pro (May 2025)Google DeepMind · Language, Vision, Video, Multimodal, SpeechAPI accessMay 6, 2025Kevin-32BCognition, Stanford University · Language · 32B parametersOpen weights (non-commercial)May 6, 2025MiniMax-Speech-02-turboMiniMax · SpeechAPI accessMay 2, 2025DeepSeek-Prover-V2-671BDeepSeek · Language · 671B parametersOpen weights (restricted use)Apr 30, 2025DeepSeek-Prover-V2-7BDeepSeek · Language · 7B parametersOpen weights (restricted use)Apr 30, 2025GTE-ModernColBERT-v1LightOn · Language · 149M parameters · Training compute 3.7e21 FLOPOpen weights (unrestricted)Apr 30, 2025Phi-4-ReasoningMicrosoft · Language · 14B parameters · Training compute 9.3e23 FLOPOpen weights (unrestricted)Apr 30, 2025Phi-4-Reasoning-plusMicrosoft · Language · 14B parametersOpen weights (unrestricted)Apr 30, 2025Qwen3-0.6BAlibaba · Language · 600M parameters · Training compute 1.3e23 FLOPOpen weights (unrestricted)Apr 29, 2025Qwen3-14BAlibaba · Language · 14.8B parameters · Training compute 3.2e24 FLOPOpen weights (unrestricted)Apr 29, 2025Qwen3-1.7BAlibaba · Language · 1.7B parameters · Training compute 3.7e23 FLOPOpen weights (unrestricted)Apr 29, 2025Qwen3-235B-A22BAlibaba · Language · 235B parameters · Training compute 4.8e24 FLOPOpen weights (unrestricted)Apr 29, 2025Qwen3-30B-A3BAlibaba · Language · 30B parameters · Training compute 6.5e23 FLOPOpen weights (unrestricted)Apr 29, 2025Qwen3-32BAlibaba · Language · 32.8B parameters · Training compute 7.1e24 FLOPOpen weights (unrestricted)Apr 29, 2025
Data: Epoch AI, CC BY 4.0. Changes: selected rows and columns of Epoch AI's database.