DeepSeek V4 Flash (0731) (High): benchmark results
Provider: DeepSeek. Released 2026-07-31. Access: Open.
Aggregate Rating 1760 ± 16, rank #224 of 2171 rated models, from 47 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Vals AI LiveCodeBench | 87.26 | Accuracy (%) | 88.7 |
| Vals AI SWE-bench Verified | 88.8 | Resolved (%) | 88.5 |
| ObviousBench | 98.61 | Answer pass³ (%) | 83.2 |
| Vals AI GPQA | 89.9 | Accuracy (%) | 82.1 |
| Chess Bench LLM | 1381 | Lichess-Classical-calibrated Glicko-2 rating (native signed | 81.8 |
| WebDev Arena | 1580.2 | Arena Score | 78.8 |
| LLMConfBench | 0.33 | Conformer Ranking Accuracy (mean reliable-pair Kendall tau) | 77.8 |
| Vals AI MMLU-Pro | 86.21 | Accuracy (%) | 67.1 |
| Vals AI Harvey Legal Agent Bench | 8.33 | Accuracy (%) | 66.9 |
| GameLogicBench - Repo | 26.09 | Solve rate (%; the 23 Repo tasks (a mechanic inside a real o | 66.7 |
| WeirdML | 57.05 | Average Score | 66.5 |
| Vals AI Vibe Code Bench | 74.74 | Accuracy (%) | 66.1 |
Interactive version: theaggregate.ai/model?slug=deepseek-v4-flash-0731-high · How It Works · Data refreshed daily, snapshot 2026-11-06.