DeepSeek V4 Pro (0813): benchmark results
Provider: DeepSeek. Released 2026-08-13. Access: Open.
Unified ELO 1711 ± 1, rank #42 of 1392 rated models, from 81 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| LLM Stats (Humanity's Last Exam (with tools, text-only)) | 60 | Score (%) | 100 |
| LLM Stats (NL2Repo) | 61.5 | Score (%) | 100 |
| LLM Stats Score | 52.49 | LLM Stats Score (conservative rating) | 97.4 |
| AI Chess Leaderboard (Continuation) | 1502 | Elo | 96.2 |
| BenchmarkList ECI | 148.12 | Capability Index (ECI) | 94.9 |
| OpenRouter Tau2-Bench Airline | 78 | Accuracy (%) | 94.1 |
| LLM Stats (Toolathlon) | 74.1 | Score (%) | 93.8 |
| AI Chess Leaderboard (Reasoning) | 1484 | Elo | 92.3 |
| NL2Repo | 61.1 | Score (self-reported) | 89.7 |
| LiveBench Consecutive Events | 90.23 | Score | 88.7 |
| Gert Labs Rankings | 59 | GScore (%) | 87.9 |
| LiveBench Connections | 100 | Score | 87.7 |
Interactive version: theaggregate.ai/model?slug=deepseek-v4-pro-0813 · How It Works · Data refreshed daily, snapshot 2026-09-05.