Mathematics benchmark
Apex leaderboard
Every model the catalog carries a published Apex value for, ranked by that value.
CategoryMathematics
MeasurePass@1 math benchmark
TasksAdvanced mathematical reasoning
DifficultyFrontier math reasoning
Published byDeepSeek-V4 Technical Report
A high-difficulty mathematical reasoning benchmark reported in DeepSeek-V4 model evaluations.
Apex ranking
8 models with a published Apex value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Pass@1 math benchmark |
|---|---|---|---|
| 1 |  Hy4 preview | Tencent | 74.2 |
| 2 |  DeepSeek V4.1 Flash | DeepSeek | 65.6 |
| 3 | STA.X K2 | SK Telecom | 45.8 |
| 4 |  Qwen3.7 Max | Alibaba | 44.5 |
| 5 | DeepSeek V4 Pro 0813 | DeepSeek | 38.3 |
| 6 | DeepSeek V4 Flash 0731 | DeepSeek | 33.0 |
| 7 | ZYZAYA1-8B | Zyphra | 32.2 |
| 8 | Qwen3.7 Plus | Alibaba | 22.7 |
Evidence key: Observed
Rows are ordered by the value DeepSeek-V4 Technical Report published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.