Models
Models from the labs in Epoch AI's catalog, the repositories they published on Hugging Face, and posts from their own news pages.
DolphinGemmaGoogle DeepMind, Georgia Institute of Technology, Wild Dolphin Project · Audio · 400M parametersUnreleasedApr 14, 2025AMIE (Articulate Medical Intelligence Explorer)Google DeepMind, Google Research · Medicine, Language · 340B parameters · Training compute 7.3e24 FLOPUnreleasedApr 9, 2025T5Gemma (Gemma 9B-9B)Google DeepMind · Language · 16.7B parametersOpen weights (restricted use)Apr 8, 2025TxGemma 27BGoogle DeepMind, Google Research · Language, Biology · 27B parameters · Training compute 2.1e24 FLOPOpen weights (restricted use)Apr 8, 2025TxGemma 2BGoogle DeepMind, Google Research · Language, Biology · 2.6B parameters · Training compute 3.2e22 FLOPOpen weights (restricted use)Apr 8, 2025TxGemma 9BGoogle DeepMind, Google Research · Language, Biology · 9B parameters · Training compute 4.4e23 FLOPOpen weights (restricted use)Apr 8, 2025Gemini 2.5 Pro (Mar 2025)Google DeepMind · Language, Vision, Video, Multimodal, SpeechAPI accessMar 25, 2025Chirp 3 HD Text-to-SpeechGoogle, Google DeepMind · SpeechAPI accessMar 17, 2025Chirp 3 Speech-to-TextGoogle, Google DeepMind · SpeechHosted access (no API)Mar 17, 2025Gemini RoboticsGoogle DeepMind · Robotics, Vision, SpeechUnreleasedMar 12, 2025Gemini Robotics-ERGoogle DeepMind · Robotics, Vision, SpeechUnreleasedMar 12, 2025Gemma 3 12BGoogle DeepMind · Language, Vision, Multimodal · 12B parameters · Training compute 8.6e23 FLOPOpen weights (restricted use)Mar 12, 2025Gemma 3 1BGoogle DeepMind · Language · 1B parameters · Training compute 1.2e22 FLOPOpen weights (restricted use)Mar 12, 2025Gemma 3 27BGoogle DeepMind · Language, Vision, Multimodal · 27B parameters · Training compute 2.3e24 FLOPOpen weights (restricted use)Mar 12, 2025Gemma 3 4BGoogle DeepMind · Language, Vision, Multimodal · 4B parameters · Training compute 9.6e22 FLOPOpen weights (restricted use)Mar 12, 2025SigLIP 2Google DeepMind · Vision · 1.1B parameters · Training compute 8.2e22 FLOPOpen weights (unrestricted)Feb 20, 2025Gemini 2.0 Flash-LiteGoogle DeepMind · Language, Vision, Video, Speech, MultimodalAPI accessFeb 5, 2025
Data: Epoch AI, CC BY 4.0. Changes: selected rows and columns of Epoch AI's database.