DeepSeek V4-Pro
DeepSeek llmDeepSeek's V4 flagship with a 1M-token context, combining Compressed Sparse Attention and Heavily Compressed Attention. Scores are from its Think Max extended-reasoning mode.
overall ⌄
68.3
11 of 93 ranked
price · $/M tokens
0.435 / 0.87
in / out, as of 2026-07-25
to run it yourself ⌄
960 GB
minimum · 3.8 TB recommended
Benchmarks
What its maker published, and what anyone else measured. Every figure links the document it came from.
Cheaper, and at least as good
Nothing is both cheaper and ranked as high. This model is on the value frontier — whatever you pay for it, you are paying for something.
Lineage
No recorded lineage — root or standalone model.
Metadata
| org | DeepSeek |
| released | 2026-04-22 |
| params | 1600B-A49B |
| license | mit |
| hf id | deepseek-ai/DeepSeek-V4-Pro |