DeepSeek V4 Flash (Reasoning): benchmark results
Provider: DeepSeek. Released 2026-04-23. Access: Open.
Unified ELO 1632 ± 1, rank #310 of 1761 rated models, from 9 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| UGI Leaderboard | 56.63 | UGI Score | 97.4 |
| UGI - Natural Intelligence | 45.52 | NatInt Score | 88.4 |
| UGI - Writing | 49.74 | Writing Score | 88.2 |
| Wolfram LLM Benchmarking Project | 59.6 | Correct Functionality (%) | 88 |
| AtmosCoder-Bench | 91.1 | Accuracy (%, mean of 3 runs) | 66.7 |
| GRIPS | 88.2 | Accuracy (%) | 61.5 |
| GRIPS-hard | 49.3 | Accuracy (%) | 61.5 |
| UGI - Willingness (W/10) | 5.2 | W/10 Score | 42 |
| SpeechMap Compliance | 50.7 | % Requests Completed | 33.5 |
Interactive version: theaggregate.ai/model?slug=deepseek-v4-flash-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-05.