DeepSeek
DeepSeek-V4-Pro benchmark results
11 results on 5 benchmarks from 2 sources.
Also listed as: deepseek-v4-pro-high-20260813 · deepseek-v4-pro · deepseek-v4-pro_none · DeepSeek v4 Pro (unknown thinking) · deepseek-v4-pro_high · DeepSeek v4 Pro (high) · deepseek-v4-pro_max · DeepSeek v4 Pro (max)
All benchmark results
| Benchmarks | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|
| LMArena Text | 1465 | 55 | Overall, style control on; 9815 votes; 95% CI 1458–1471“deepseek-v4-pro-high-20260813” | Official leaderboard | Oct 8, 2026 |
| LMArena Text | 1458 | 62 | Overall, style control on; 57657 votes; 95% CI 1454–1462“deepseek-v4-pro” | Official leaderboard | Oct 8, 2026 |
| LMArena WebDev | 1582 | 32 | Overall; 4740 votes; 95% CI 1573–1591“deepseek-v4-pro-high-20260813” | Official leaderboard | Oct 8, 2026 |
| LMArena WebDev | 1447 | 70 | Overall; 14019 votes; 95% CI 1440–1453“deepseek-v4-pro” | Official leaderboard | Oct 8, 2026 |
| GPQA Diamond | 90.9% | 31 | Epoch AI run, accuracy; effort: high | Independent run | Aug 6, 2026 |
| GPQA Diamond | 89.7% | 47 | Epoch AI run, accuracy; effort: max | Independent run | Jun 16, 2026 |
| GPQA Diamond | 73.2% | 148 | Epoch AI run, accuracy; effort: unknown thinking | Independent run | Aug 6, 2026 |
| SWE-bench Verified | 77.6% | 6 | Epoch AI run (Epoch scaffold); effort: max | Independent run | Jun 18, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 96.7% | 39 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: max | Independent run | Jun 17, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 95.6% | 43 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: high | Independent run | Aug 6, 2026 |
| AIME (OTIS Mock AIME 2024–2025) | 46.7% | 172 | Epoch AI run, OTIS Mock AIME 2024–2025; effort: unknown thinking | Independent run | Aug 6, 2026 |
Related on this site
FAQ
Where do these DeepSeek-V4-Pro scores come from?
From Epoch AI Benchmarking Hub and LMArena leaderboard dataset. Each row links to the original leaderboard, dataset or report, and vendor-reported numbers are labelled.