The Model Index
Models/DeepSeek

DeepSeek-V3.1

DeepSeekReleased 21 Aug 2025Open
CompareAnnouncement
132120–144Index score#111 · provisional, limited data
0/ 8Core tests taken0 of 6 areas · plus 5 anchor tests
−22Versus the frontier at releasebest then: GPT-5Replaced by DeepSeek-V3.2-Exp
$0.25/ $0.95Price per million tokensinput / output, list price
164KContext windowtokens it reads at once

Where it sitson the frontier

o1 reached this score 9 months earlier.

Test resultsby area · tick = best by any model

Reasoning & knowledgeevidence
GPQA Diamond
no result
Humanity's Last Exam
no result
Mathsno evidence yet
FrontierMath T1-3
no result
Codingevidence
SWE-bench Verified
no result
Terminal-Bench
no result
Agentsevidence
APEX-Agents
no result
Novel problemsno evidence yet
ARC-AGI-2
no result
Long tasksno evidence yet
METR time horizon
no result

Independent results from the Epoch AI benchmarking hub, best reasoning setting. The core tests are shown here; anchor tests also inform the score and can be shown above.

What others sayrank on each leaderboard

The Model Indexindependent tests, our method#111of 219 · 132
LMArena · Creative writingcreative writing#93of 386 · 1402LMArena · Textoverall chat quality, judged by people#116of 388 · 1417LMArena · Mathsmaths questions in chat#120of 373 · 1414LMArena · Codingcoding questions in chat#130of 383 · 1456Epoch Capabilities Indexoverall capability from many benchmarks#105of 239 · 139.9

Further left is better. Leaderboards measure different things: LMArena is people voting blind between two answers, the others are test scores. Click any row for the source.

Best attop three among current models

Replaced by DeepSeek-V3.2-Exp; it no longer leads any category among current models.

Where this model places in the top three of today's models, on category leaderboards and on our core tests.

Where to use itprices checked 1 Oct 2026

From the lab

DeepSeek Chatapp · subscription↗DeepSeek APIAPI↗

API providers 7 · per million tokens, in / out

DDeepInfra$0.25 / $0.95 · fp4cheapestSSiliconFlow$0.27 / $1.0 · fp8↗AAtlasCloud$0.30 / $1.0 · fp8↗CCoreWeave$0.55 / $1.65 · fp8↗SSambaNova$0.65 / $1.5 · fp8↗MMara$0.60 / $1.7↗GGoogle$0.60 / $1.7↗

Run it yourself

🤗Official weightsdeepseek-ai/DeepSeek-V3.1↗

Provider prices come from OpenRouter's live listing; going direct to a provider can differ. Subscriptions are the lab's own apps.

Facts and sources

Released21 Aug 2025
Release entryDeepSeek-V3.1 (hybrid reasoning)
Scored asDeepSeek-V3.1
Artificial Analysis index–
Output speed–

DeepSeek releasesaround this one

DeepSeek-R1Open
index 129
DeepSeek-R1-0528Open
index 134
DeepSeek-V3.1Open
index 132
DeepSeek-V3.2-ExpOpen
index 143
DeepSeekMath-V2Open