DeepSeek-V4-Pro
Key facts
- Released
- Apr 24, 2026
- What it does
- Text and chat: Writes, answers questions and holds a conversation.Thinks first: Can reason step by step before it answers: slower, but better at maths, code and tricky questions.
View moreView less
- How
- Free to download
- Size
- Parameters: The numbers a model learns during training: a rough measure of its size. Bigger usually means more capable, but slower and pricier to run. 7B means 7 billion.
- Announced as
- “DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence”
From trckr's archive of releases before it started watching. Date and details from Epoch AI's database of AI models (CC BY 4.0).
Price and limits
What developers pay to use it through the API: How developers plug a model into their own apps and pay per use. A model that's "in the API" can be used by programmers, even if it isn't in a chat app yet., per million Tokens: The small chunks of text a model reads and writes, about three-quarters of a word each. Prices and limits are counted in tokens. sent (input) or written back (output), and how much it can take in at once.
- Input
- $0.66
- per million tokensCached: $0.022
- Output
- $1.98
- per million tokens
- Context window: How much a model can keep in mind at once: your question, any files you share and its own replies. Measured in tokens; 1,000 tokens is about 750 words.
- 1.05M tokens
- about 790,000 words
View moreView less
- Longest answer
- 393K tokens (about 290,000 words)
From the lab's own listing on models.dev and OpenRouter, checked .
How it scores
From independent testers, never the lab's own claims.
- Combined test score#36 of 235
- GPT-6 Astra169
- 2GPT-6.1 Sol169
- 35Grok 4.7156
- 36DeepSeek V4 Pro 0813156
- 37Qwen 3.8 Max156
Source: Epoch AI · Epoch Capabilities Index
- Chat answers, by people's votes#50 of 86
- Gemini 4 Carbon1,527
- 2Gemini 4 Argon1,522
- 49Qwen3.5-Max-Preview1,465
- 50DeepSeek V4 Pro 08131,464
- 51Claude Sonnet 51,461
Source: LMArena · LMArena text, adjusted for style
- Building apps, by people's votes#37 of 85
- Claude Opus 5.51,818
- 2Claude Sonnet 5.51,773
- 36DeepSeek V4 Flash 07311,582
- 37DeepSeek V4 Pro 08131,581
- 38GPT-6 Luna1,577
Source: LMArena · LMArena WebDev
Compare DeepSeek-V4-Pro with
It came out 148 days after DeepSeekMath-V2, the lab's previous model.
The details
Get alerts like this the moment they happen.
From DeepSeek and every other major lab.