DeepSeek V4 Pro (0813) (High): benchmark results
Provider: DeepSeek. Released 2026-08-13. Access: Open.
Aggregate Rating 1799 ± 17, rank #142 of 2170 rated models, from 59 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Chatbot Arena (Text - Korean) | 1464 | Arena Score | 95.5 |
| Chatbot Arena (Text - Legal & Government) | 1480 | Arena Score | 90.4 |
| Chatbot Arena (Text - English) | 1479 | Arena Score | 90.2 |
| Chatbot Arena (Text - Life, Physical & Social Science) | 1487 | Arena Score | 89.2 |
| Chatbot Arena (Text - Writing, Literature & Language) | 1454 | Arena Score | 89 |
| Chatbot Arena (Text - Business, Management & Financial Ops) | 1466 | Arena Score | 88.8 |
| Chatbot Arena (Text - Instruction Following) | 1461 | Arena Score | 88.4 |
| Chatbot Arena (Text - Creative Writing) | 1445 | Arena Score | 88.1 |
| GameLogicBench - Repo | 34.78 | Solve rate (%; the 23 Repo tasks (a mechanic inside a real o | 87.5 |
| Chatbot Arena (Text) | 1465 | Elo | 87.3 |
| Chatbot Arena (Text - Hard Prompts (English)) | 1489 | Arena Score | 87 |
| Chatbot Arena (Text - Multi-Turn) | 1472 | Arena Score | 87 |
Interactive version: theaggregate.ai/model?slug=deepseek-v4-pro-0813-high · How It Works · Data refreshed daily, snapshot 2026-11-05.