AI Models from DeepSeek

DeepSeek-R1 Groundbreaking open reasoning model — rivals OpenAI o1 on math, science, and cod... DeepSeek-R1 (May 28) May 2025 update to DeepSeek-R1 — improved reasoning depth, better code generatio... DeepSeek-V3 DeepSeek flagship chat & code model — top-tier on coding benchmarks, long contex... DeepSeek-V3 (Mar 24) March 2024 DeepSeek-V3 checkpoint — improved coding and reasoning over the origi... DeepSeek-V3.1 DeepSeek-V3.1 — next iteration with improved instruction following, better codin... DeepSeek-V3.1 Terminus Terminus edition of DeepSeek-V3.1 — enhanced for agentic workflows, tool use, an... DeepSeek-V3.2 Exp Experimental V3.2 — leading SWE-bench scores. Frontier-level coding agent. Not s... DeepSeek-R1 Distill Qwen 1.5B Tiny R1 distillation — astonishing reasoning at 1.5B params. Perfect for local d... DeepSeek-R1 Distill Qwen 7B Fast chain-of-thought reasoning in a compact 7B package. Excellent math, logic, ... DeepSeek-R1 Distill Qwen 32B Powerful R1 distillation in a 32B package — matches GPT-4-class performance on m... DeepSeek-R1 Distill Llama 8B Llama-based R1 distillation — fast reasoning using Llama-3 architecture backbone... DeepSeek-R1 Distill Llama 70B Large Llama-based R1 distillation — powerful reasoning across math, science, cod... DeepSeek-R1 (Thinking Model) DeepSeek-R1 extended thinking model — chain-of-thought reasoning for math, scien... DeepSeek V4 Flash DeepSeek V4 Flash — ultra-fast variant of DeepSeek V4 for quick reasoning and co... DeepSeek V4 Pro DeepSeek V4 Pro — professional-grade DeepSeek model with enhanced reasoning, cod...