> DeepSeek-V3 is an open-weights AI model from DeepSeek, released 26 Dec 2024. Index score 116 (GPT-4 = 100), rank 145 of 229 ranked models. API price $0.26 in / $1.03 out per million tokens, 164K context.

[Models](https://themodelindex.org/models/) /[DeepSeek](https://themodelindex.org/labs/deepseek/)

# DeepSeek-V3

DeepSeek

Released 26 Dec 2024

Open

[Compare](https://themodelindex.org/compare/deepseek-v3-vs-o1/) [Announcement](https://api-docs.deepseek.com/news/news1226/)

116105–128

Index score

#145 of 229 models

2/ 8

Core tests taken

2 of 6 areas · plus 6 anchor tests

−20

Versus the frontier at release

best then: o1

Replaced by [DeepSeek-V3 (Mar 2025)](https://themodelindex.org/models/deepseek-v3-mar-2025/)

$0.26/ $1.03

Price per million tokens

input / output, list price

164K

Context window

tokens it reads at once

## Where it sits · on the frontier

_Chart: Model Index score of every scored model since late 2022, with the best closed and best open model over time_

o1-mini reached this score 3 months earlier.

## Test results · by area · tick = best by any model

Show 6 anchor tests

Reasoning & knowledge

evidence

GPQA Diamond

57%

Humanity's Last Exam

no result

Maths

evidence

FrontierMath T1-3

no result

Coding

evidence

SWE-bench Verified

no result

Terminal-Bench

no result

Agents

no evidence yet

APEX-Agents

no result

Novel problems

no evidence yet

ARC-AGI-2

no result

Long tasks

evidence

METR time horizon

18 min

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 say · rank on each leaderboard

The Model Index

independent tests, our method

#145

of 219 · 116

LMArena · Creative writing

creative writing

#159

of 386 · 1349

LMArena · Text

overall chat quality, judged by people

#183

of 388 · 1358

Epoch Capabilities Index

overall capability from many benchmarks

#118

of 239 · 135.9

LMArena · Coding

coding questions in chat

#200

of 383 · 1387

LMArena · Maths

maths questions in chat

#213

of 373 · 1310

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 at · top three among current models

Replaced by DeepSeek-V3 (Mar 2025); 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 it · prices checked 1 Oct 2026

#### API providers · 2 · per million tokens, in / out

S

StreamLake

$0.26 / $1.03

cheapest

D

DeepInfra

$0.32 / $0.89 · fp4

↗

#### Run it yourself

🤗

Official weights

deepseek-ai/DeepSeek-V3

↗

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

Released

26 Dec 2024

Announcement

[api-docs.deepseek.com ↗](https://api-docs.deepseek.com/news/news1226/)

Release entry

DeepSeek-V3

Scored as

DeepSeek-V3

Artificial Analysis index

–

Output speed

–

Hugging Face

[deepseek-ai/DeepSeek-V3 ↗](https://huggingface.co/deepseek-ai/DeepSeek-V3)

## DeepSeek releases · around this one

DeepSeek-Coder-V2

Open

17 JUN 2024

DeepSeek-R1-Lite-Preview

Closed

20 NOV 2024

DeepSeek-V3

Open

26 DEC 2024

index 116

DeepSeek-R1

Open

20 JAN 2025

index 129

DeepSeek-R1-0528

Open

28 MAY 2025

index 134

---
Source: https://themodelindex.org/models/deepseek-v3/ · The Model Index · data as of 2026-10-01