DeepSeek-V3.1
DeepSeek-V3.1 is an open-weights AI model from DeepSeek with an index score of 128 (GPT-4 = 100), rank 136 of 217.
128111–145Index score#136 of 217 models
$0.25/ $0.95Price per million tokensinput / output · via OpenRouter
164KContext windowtokens it reads at once
21 Aug 2025ReleasedReplaced by DeepSeek-V3.2-Exp
Where it sitson the frontier
o1 reached this score 9 months earlier.
Test resultsby area · tick = best by any model
Reasoning & knowledgeevidence
Mathsno evidence yet
Codingevidence
Agentsevidence
Professional workevidence
Novel problemsno evidence yet
Long tasksno evidence yet
Third-party results only, at the model’s best standard setting. Hover a bar for who ran it.
What others sayrank on 8 leaderboards
The Model Indexthird-party tests, our method#136of 217 · 128
LMArena · Creative writingcreative writing#88of 360 · 1401LMArena · Long promptslong, complex requests#84of 340 · 1446LMArena · Instruction followingfollowing detailed instructions#98of 362 · 1416LMArena · Textoverall chat quality, judged by people#112of 362 · 1417LMArena · Mathsmaths questions in chat#116of 346 · 1414LMArena · Codingcoding questions in chat#128of 357 · 1455Epoch Capabilities Indexoverall capability from many benchmarks#115of 231 · 140.8Best attop three among current models
Replaced by DeepSeek-V3.2-Exp; it no longer leads any category among current models.
Where to use itprices checked 18 Nov 2026
API providers 3 · per 1M tokens, in / out
Run it yourself
Official weightsdeepseek-ai/DeepSeek-V3.1↗Host prices from OpenRouter; going direct can differ.
Facts and sources
Released21 Aug 2025
Announcementapi-docs.deepseek.com ↗
Release entryDeepSeek-V3.1 (hybrid reasoning)
Scored asDeepSeek-V3.1
Output speed–
Hugging Facedeepseek-ai/DeepSeek-V3.1 ↗