DeepSeek V4 Flash (0423): benchmark results
Provider: DeepSeek. Released 2026-04-24. Access: Open.
Aggregate Rating 1698 ± 18, rank #359 of 10916 rated models, from 171 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Evals for Every Language - Romanian (ro) | 80.17 | Average Score (%) | 100 |
| Evals for Every Language - Lao (lo) | 76.13 | Average Score (%) | 98.5 |
| Evals for Every Language - Yoruba (yo) | 69.78 | Average Score (%) | 97 |
| Evals for Every Language - Kannada (kn) | 75.98 | Average Score (%) | 96.9 |
| Evals for Every Language - Māori (mi) | 74.32 | Average Score (%) | 95.4 |
| Evals for Every Language - Armenian (hy) | 74.73 | Average Score (%) | 95.3 |
| Evals for Every Language - Sanskrit (sa) | 70.78 | Average Score (%) | 95.2 |
| Evals for Every Language - Burmese (my) | 72.59 | Average Score (%) | 95.1 |
| Evals for Every Language - Amharic (am) | 73.72 | Average Score (%) | 95 |
| Evals for Every Language - Arabic (ar) | 72.7 | Average Score (%) | 93.8 |
| Evals for Every Language - Central Kurdish (ckb) | 72.94 | Average Score (%) | 93.8 |
| Evals for Every Language - Crimean Tatar (crh) | 72.1 | Average Score (%) | 93.8 |
Interactive version: theaggregate.ai/model?slug=deepseek-v4-flash-0423 · How It Works · Data refreshed daily, snapshot 2026-11-08.