Models
Models from the labs in Epoch AI's catalog, the repositories they published on Hugging Face, and posts from their own news pages.
DeepSeek V4.1 Flash 1017DeepSeek · Language, Multimodal, VisionOpen weights (unrestricted)Oct 17DeepSeek V4.1 FlashDeepSeek · Language, Vision, Multimodal · 748B parameters · Training compute 4.3e24 FLOPOpen weights (unrestricted)Sep 9DeepSeek V4 Pro 0813DeepSeek · Language · 1,600B parameters · Training compute 9.7e24 FLOPOpen weights (unrestricted)Aug 13DeepSeek V4 Flash 0731DeepSeek · Language · 284B parameters · Training compute 2.5e24 FLOPOpen weights (unrestricted)Jul 31DeepSeek-V4-FlashDeepSeek · Language · 284B parameters · Training compute 2.5e24 FLOPOpen weights (unrestricted)Apr 24DeepSeek-V4-ProDeepSeek · Language · 1,600B parameters · Training compute 9.7e24 FLOPOpen weights (unrestricted)Apr 24DeepSeek-V3.2DeepSeek · Training compute 4.2e24 FLOPOpen weights (unrestricted)Dec 1, 2025DeepSeekMath-V2DeepSeek · Language · 685B parametersOpen weights (unrestricted)Nov 27, 2025Deepseek OCRDeepSeek · Vision, Language · 3B parametersOpen weights (unrestricted)Oct 21, 2025DeepSeek-V3.2-ExpDeepSeek · Language · 671B parameters · Training compute 4.2e24 FLOPOpen weights (unrestricted)Sep 29, 2025DeepSeek-V3.1-TerminusDeepSeek · Language · 671B parameters · Training compute 3.6e24 FLOPOpen weights (unrestricted)Sep 22, 2025DeepSeek-V3.1DeepSeek · Language · 671B parameters · Training compute 3.6e24 FLOPOpen weights (unrestricted)Aug 21, 2025DeepSeek-R1 (May 2025)DeepSeek · Language · 671B parameters · Training compute 4e24 FLOPOpen weights (unrestricted)May 28, 2025DeepSeek-Prover-V2-671BDeepSeek · Language · 671B parametersOpen weights (restricted use)Apr 30, 2025DeepSeek-Prover-V2-7BDeepSeek · Language · 7B parametersOpen weights (restricted use)Apr 30, 2025DeepSeek-V3 (Mar 2025)DeepSeek · Language · 671B parameters · Training compute 3.3e24 FLOPOpen weights (restricted use)Mar 24, 2025Janus-Pro-1BDeepSeek · Image generation, Vision, Language, Multimodal · 1B parametersOpen weights (restricted use)Jan 27, 2025Janus-Pro-7BDeepSeek · Image generation, Vision, Language, Multimodal · 7B parametersOpen weights (restricted use)Jan 27, 2025DeepSeek-R1-Distill-Llama-70BDeepSeek · Language · 70B parametersOpen weights (unrestricted)Jan 22, 2025DeepSeek-R1-Distill-Llama-8BDeepSeek · Language · 8B parametersOpen weights (restricted use)Jan 22, 2025DeepSeek-R1-Distill-Qwen-14BDeepSeek · Language · 14.8B parametersOpen weights (unrestricted)Jan 22, 2025DeepSeek-R1-Distill-Qwen-1.5BDeepSeek · Language · 1.8B parametersOpen weights (unrestricted)Jan 22, 2025DeepSeek-R1-Distill-Qwen-32BDeepSeek · Language · 32B parametersOpen weights (unrestricted)Jan 22, 2025DeepSeek-R1-Distill-Qwen-7BDeepSeek · Language · 7B parametersOpen weights (unrestricted)Jan 22, 2025DeepSeek-R1DeepSeek · Language · 671B parameters · Training compute 3.5e24 FLOPOpen weights (unrestricted)Jan 20, 2025
Data: Epoch AI, CC BY 4.0. Changes: selected rows and columns of Epoch AI's database.