LIVETracking 1,072 AI models from 273 providers.
Beta
Home/Models/DeepSeek V4 Pro
DeepSeek logo

DeepSeek V4 Pro

by DeepSeek · Released Apr 2026

Open Source1M Context
86.5
avg score
Rank #11
Compare
Top 4% of all models
Context
1.0M tokens (~524 books)
Input $/1M
$0.96
Output $/1M
$1.91
Type
text
License
Open Source
Benchmarks
12 tested
Data updated today
About

DeepSeek V4 Pro is a large-scale Mixture-of-Experts model from DeepSeek with 1.6T total parameters and 49B activated parameters, supporting a 1M-token context window. It is designed for advanced reasoning, coding,...

Tested on 12 benchmarks with 73.6% average. Top scores: Chatbot Arena Elo — Overall (1455.9%), LiveBench — Mathematics (90.7%), LiveBench — Reasoning (82.7%).

Capabilities
coding
63.3
#39 globally
reasoning
78.6
#14 globally
math
90.7
#8 globally
knowledge
73.6
#12 globally
language
70.2
#75 globally
speed
77.8
#15 globally
Benchmark Scores
Compare All
Tested on 12 benchmarks · Ranked across 7 categories
Score Distribution (all 274 models)
0255075100
▲ You are here
codingCompare coding →
LiveBench — Coding

Regularly refreshed coding problems that avoid data contamination. New problems added monthly to prevent memorization.

70.0—
LiveBench — Agentic Coding

LiveBench coding tasks that require multi-step reasoning and tool use. Tests planning and execution of complex coding workflows.

56.7—
reasoningCompare reasoning →
LiveBench — Reasoning

Regularly refreshed reasoning problems testing logical deduction, spatial reasoning, and analytical thinking.

82.7—
LiveBench — Data Analysis

Fresh data analysis tasks testing ability to interpret tables, charts, and statistical data.

74.5—
mathCompare math →
LiveBench — Mathematics

Regularly updated math problems that test numerical reasoning, algebra, calculus, and combinatorics.

90.7—
Quick compare:DeepSeek logovs DeepSeek-V2 (MoE-236B, May 2024)OpenAI logovs GPT-5.1-Codex-Max
HA
vs Qwen2.5 72B Instruct Abliterated
Excellent (85+) Good (70-85) Average (50-70) Below (<50)
Model Family · DeepSeek DeepSeek V4
Apr 2026
67.2
$0.08/M in1.0Mctx12 benchmarks
Apr 2026
73.6+6.4
$0.43/M in(+0.35)1.0Mctx12 benchmarks
See the full DeepSeek V4 family →
Similar Models
DeepSeek logo
DeepSeek-V2 (MoE-236B, May 2024)
DeepSeek
86.1TBD
OpenAI logo
GPT-5.1-Codex-Max
OpenAI
85.6$1.25/1M
HA
Qwen2.5 72B Instruct Abliterated
HuiHui AI
87.5TBD
Links
Info
DeepSeekPricing explorerDevelopers · API
Research
Technical Report
Documentation
API DocsPlayground
Community
@DeepSeek
Source Code
GitHubHugging Face
BenchGecko API
deepseek-v4-pro
Specifications
  • Typetext
  • Context1.0M tokens (~524 books)
  • ReleasedApr 2026
  • LicenseOpen Source
  • StatusActive
  • Cost / Message~$0.004
Available On
DeepSeek logoDeepSeek$0.96
Categories
codingreasoningmathknowledgelanguagespeed
Learn More
context-windowtransformeropen weightstokens
Share & Export
Tweet
Related Models
DeepSeek logoDeepSeek-V2 (MoE-236B, May 2024)
86.1
OpenAI logoGPT-5.1-Codex-Max
85.6
HA
Qwen2.5 72B Instruct Abliterated
87.5
DeepSeek logoDeepSeek R1 Distill Qwen 14B
85.4
DeepSeek V4 Pro is an open-source text AI model by DeepSeek, released in April 2026. It has an average benchmark score of 86.5. Context window: 1M tokens.

Related Models

DeepSeek-V2 (MoE-236B, May 2024) · DeepSeekGPT-5.1-Codex-Max · OpenAIQwen2.5 72B Instruct Abliterated · HuiHui AIDeepSeek R1 Distill Qwen 14B · DeepSeekGPT-5.4 Pro · OpenAI

Benchmarks

Chatbot Arena Elo — OverallLiveBench — MathematicsLiveBench — ReasoningLiveBench — LanguageLiveBench — Data Analysis

Related Pages

DeepSeek · ProviderDeepSeek · EconomyAll ModelsCompare ModelsPricingDevelopers · APIContext Window · Glossary