Skip to content
VECTOR WIREAI INTELLIGENCE
⌘K
UTC
Models741Deals30d268Policy30d67, 15 pendingSources1,014

DeepSeek-V3

DeepSeekReleased Dec 26, 2024Current

Score basis

DeepSeek-V3 is behind the leaders in factuality and coding. Too few results yet to rate reasoning, agentic tasks, safety, long context, math, multimodal tasks, multilingual tasks, or instruction following.

Capability profile

Bars show the model's median result as a share of the leading model's, per capability. Select a row to see its results.

Limited2 capabilities
  1. Factuality−31.3%3 of 4
  2. Coding−44.6%5 of 10
Not rated8 capabilities

Too few results yet to rate Reasoning, Agentic, Safety, Long Context, Math, Multimodal, Multilingual or Instruction Following.

Price

$0.24input$0.90outputper million tokens

From Artificial Analysis · 3 providers tracked · All prices

Cheaper than 65% of 331 priced models · 3:1 input-to-output blend, log scale

Evidence

205results on131benchmarks

  • 35 independently verified
  • 36 aggregator
  • 91 vendor-reported
  • 43 cross-referenced

From 30 sources · latest Oct 8, 2026 · How verification works

API features

Tool callingStructured outputs

As listed by OpenRouter

Research

72 papers reference DeepSeek-V3

DeepSeek-V3 benchmark results

205 results on 131 benchmarks, grouped by capability. Each bar is the result as a share of the capability leader's; every result links to its source.

Factuality

LimitedFull factuality ranking

31.3% behind the leader3 of 4 ranked benchmarks measured

Show 2 more factuality resultsHide 2 factuality results

Coding

LimitedFull coding ranking

44.6% behind the leader5 of 10 ranked benchmarks measured

Show 10 more coding resultsHide 10 coding results

Reasoning

Not enough dataFull reasoning ranking

0 of 6 ranked benchmarks measured

Show 10 more reasoning resultsHide 10 reasoning results

Agentic

Not enough dataFull agentic ranking

0 of 7 ranked benchmarks measured

Show 3 more agentic resultsHide 3 agentic results

Long Context

Not enough dataFull long context ranking

0 of 3 ranked benchmarks measured

Instruction Following

Not enough dataFull instruction following ranking

0 of 3 ranked benchmarks measured

Show 2 more instruction following resultsHide 2 instruction following results

More results

Benchmarks outside the capability baskets. They are not ranked against a leader.

Show 152 more resultsHide 152 results

DeepSeek-V3: common questions

Who makes DeepSeek-V3?

DeepSeek-V3 is made by DeepSeek.

When was DeepSeek-V3 released?

DeepSeek-V3 was released on Dec 26, 2024, according to Artificial Analysis.

What is DeepSeek-V3 good at?

DeepSeek-V3 is behind the leaders in factuality and coding. Too few results yet to rate reasoning, agentic tasks, safety, long context, math, multimodal tasks, multilingual tasks, or instruction following.

How much does DeepSeek-V3 cost?

DeepSeek-V3 costs $0.24 per million input tokens and $0.90 per million output tokens, according to Artificial Analysis. We track its price at 3 providers. At a mix of three input tokens to one output token, it is cheaper than 65% of the 331 priced models we track.

How many benchmarks has DeepSeek-V3 been tested on?

We track 205 results for DeepSeek-V3 on 131 benchmarks from 30 sources, 35 of them independently verified. The latest was recorded on Oct 8, 2026.

Which API features does DeepSeek-V3 support?

OpenRouter lists tool calling and structured outputs for DeepSeek-V3.

About this record

Where DeepSeek-V3's numbers come from, and every name it appears under.

Tracked since
Apr 25, 2026
Newest source mention
Jul 19, 2026

Where the results come from

Verification: 205 scores · 35 independently verified · 36 aggregator-attributed · 43 vendor cross-reference · 91 vendor-reported. How these tiers are assigned

From 30 sources on 10 sites. Hugging Face supplies 64 of them; the 35 independently verified results come from 9 sites. Bars are coloured by trust tier.

  • huggingface.co64
  • raw.githubusercontent.com57
  • artificialanalysis.ai37
  • api.llm-stats.com18
  • storage.googleapis.com12
  • arxiv.org6
  • livecodebench.github.io4
  • matharena.ai4
  • simple-bench.com2
  • aider.chat1

Also known as

deepseek v3 (0324)deepseek v3 (dec '24)deepseek v3 (dec)deepseek v3 03-24deepseek v3 0324deepseek-chat-v3-0324deepseek-v3 (Non-Reasoning)deepseek-v3 (Reasoning)v3