Flash: 284B total / 13B active (MoE); Pro: 1.6T total / 49B active (MoE)
1,000,000 tokens
MIT
Official DeepSeek-V4 preview release with open weights for both Flash and Pro tracks, each provided in Base and Instruct variants. Release materials emphasize million-token context efficiency via hybrid attention upgrades, stronger long-context reasoning, and expanded coding/agent capability. DeepSeek positions V4-Pro-Max as its frontier open-source reasoning mode and claims top-tier open-source performance with narrowed gaps to leading closed models on difficult reasoning and agent benchmarks. Model pages provide local deployment guidance and multi-mode reasoning controls (Non-think, Think, Think Max).