DeepSeek-V3
DeepSeek-V3 is an open-weights AI model from DeepSeek with an index score of 119 (GPT-4 = 100), rank 158 of 217.
119103–135Index score#158 of 217 models
$0.26/ $1.03Price per million tokensinput / output · via OpenRouter
164KContext windowtokens it reads at once
26 Dec 2024ReleasedReplaced by DeepSeek-V3 (Mar 2025)
Where it sitson the frontier
o1-mini reached this score 3 months earlier.
Test resultsby area · tick = best by any model
Reasoning & knowledgeevidence
Mathsevidence
Codingevidence
Agentsno evidence yet
Professional workno evidence yet
Novel problemsno evidence yet
Long tasksevidence
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#158of 217 · 119
LMArena · Creative writingcreative writing#105of 360 · 1390LMArena · Textoverall chat quality, judged by people#136of 362 · 1395LMArena · Instruction followingfollowing detailed instructions#144of 362 · 1378LMArena · Long promptslong, complex requests#144of 340 · 1394LMArena · Codingcoding questions in chat#154of 357 · 1428LMArena · Mathsmaths questions in chat#162of 346 · 1366Epoch Capabilities Indexoverall capability from many benchmarks#130of 231 · 135.3Best attop three among current models
Replaced by DeepSeek-V3 (Mar 2025); it no longer leads any category among current models.
Where to use itprices checked 18 Nov 2026
API providers 2 · per 1M tokens, in / out
Run it yourself
Official weightsdeepseek-ai/DeepSeek-V3↗Host prices from OpenRouter; going direct can differ.
Facts and sources
Released26 Dec 2024
Announcementapi-docs.deepseek.com ↗
Release entryDeepSeek-V3
Scored asDeepSeek-V3
Output speed–
Hugging Facedeepseek-ai/DeepSeek-V3 ↗