DeepSeek V4 Flash (Non-reasoning) — benchmark results
DeepSeek V4 Flash non-reasoning evaluation mode. Provider: DeepSeek. Released 2026-04-23. Access: Open.
Unified ELO 1707 ± 36, rank #262 of 1776 rated models, from 36 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| UGI Leaderboard | 59.21 | UGI Score | 98.5 |
| AA TAU-2 Bench | 94.44 | Accuracy (%) | 92.9 |
| UGI - Writing | 54.58 | Writing Score | 91.3 |
| UGI - Natural Intelligence | 47.94 | NatInt Score | 90.3 |
| AA Terminal-Bench Hard | 34.09 | Accuracy (%) | 78.8 |
| AA Omniscience - Software Engineering (SWE) - Kotlin | 32 | Accuracy (%) | 76.1 |
| AA Omniscience - Software Engineering (SWE) - Dart | 32 | Accuracy (%) | 75.8 |
| AA Omniscience - Software Engineering (SWE) - C | 56 | Accuracy (%) | 75.7 |
| AA Omniscience - Law | 20 | Accuracy (%) | 75.2 |
| AA Omniscience - Software Engineering (SWE) - JavaScript | 42.73 | Accuracy (%) | 74.7 |
| AA Omniscience - Business | 23.1 | Accuracy (%) | 74.4 |
| AA Omniscience - Software Engineering (SWE) - Go | 30 | Accuracy (%) | 74.3 |
Interactive version: theaggregate.ai/model?slug=deepseek-v4-flash-non-reasoning · How the rankings work · Data refreshed daily, snapshot 2026-07-22.