Model card
DeepSeek-V4-Pro Max.
DeepSeekLarge language model2 current SOTA
Added from Papers with Code MMLU-Pro refresh on 2026-05-19.
§ 02 · Benchmarks
Every benchmark DeepSeek-V4-Pro Max has a recorded score for.
| # | Benchmark | Area · Task | Metric | Value | Rank | Date | Source |
|---|---|---|---|---|---|---|---|
| 01 | BrowseComp | Natural Language Processing · Question Answering | accuracy | 83.4% | #1 | — | source ↗ |
| 02 | LiveCodeBench | Computer Code · Code Generation | pass-1 | 93.5% | #1 | — | source ↗ |
| 03 | SWE-Bench Verified | Computer Code · Code Generation | accuracy | 80.6% | #1 | — | source ↗ |
| 04 | MMLU-Pro | Reasoning · Commonsense Reasoning | accuracy | 87.5% | #3 | 2026-04-24 | source ↗ |
| 05 | MMLU-Pro | Reasoning · Commonsense Reasoning | accuracy | 87.5% | #3 | — | source ↗ |
| 06 | HLE | Reasoning · Multi-step Reasoning | accuracy | 37.7% | #4 | — | source ↗ |
| 07 | GPQA Diamond | Reasoning · Multi-step Reasoning | accuracy | 90.1% | #5 | — | source ↗ |
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 03 · Strengths by area
Where DeepSeek-V4-Pro Max actually performs.
§ 04 · Papers
1 paper with results for DeepSeek-V4-Pro Max.
- 2026-04-24· 6 results
DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence
§ 05 · Related models
Other DeepSeek models scored on Codesota.
§ 06 · Sources & freshness
Where these numbers come from.
pwc-dump
6
results
paperswithcode
1
result
0 of 7 rows marked verified.