DeepSeek: DeepSeek V4 Pro
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,...
Everything else
| Max output | 384K |
|---|---|
| Input / 1M | $0.435 |
| Output / 1M | $0.87 |
| Cache read / 1M | $0.004 |
| Cache write / 1M | — |
| 1M in + 1M out | $1.30 |
| Takes | text |
| Returns | text |
| Tools | yes |
| Reasoning | optional · xhigh, high |
| Open weights | yes |
| Released | 2026-04-24 |
| Knowledge cutoff | — |
| Retires | — |
| Catalog id | deepseek/deepseek-v4-pro |
Benchmarks
| source | epoch |
|---|---|
| eci | 148.71 |
| confidence | Likely |
| accessibility | Open weights (unrestricted) |
| gpqa | 0.896 |
| gpqaStderr | 0.0175 |
| sweBenchVerified | 0.776 |
| sweBenchVerifiedStderr | 0.019 |
| aime | 0.967 |
| aimeStderr | 0.02 |
| simpleQA | 0.57 |
| simpleQAStderr | 0.0157 |
Benchmark scores by Epoch AI, CC BY 4.0.
Using it
In 00, this model is picked per agent — and per tier, so the model that answers your customers need not be the one that writes your code. Get 00.