Mathematics · Benchmark profile
A high-difficulty mathematical reasoning benchmark reported in DeepSeek-V4 model evaluations.
Data verified 21 Jul 2026 · Methodology 1.6.0
Benchmark score on Apex
Qwen3.7-Max leads at 44.5%, followed by DeepSeek V4 Pro (Max) (38.3%) and DeepSeek V4 Flash (Max) (33%).
Visual analysis
Switch between model placement, score distribution and descriptive provider averages. Every view uses the same sourced leaderboard.
9 ranked models · higher is better
| Model | Rank | Provider | Score |
|---|---|---|---|
| Qwen3.7-Max | #1 | Alibaba Cloud | 44.5% |
| DeepSeek V4 Pro (Max) | #2 | DeepSeek | 38.3% |
| DeepSeek V4 Flash (Max) | #3 | DeepSeek | 33% |
| ZAYA1-8B | #4 | Zyphra | 32.2% |
| DeepSeek V4 Pro (High) | #5 | DeepSeek | 27.4% |
| Qwen3.7-Plus | #6 | Alibaba Cloud | 22.7% |
| DeepSeek V4 Flash (High) | #7 | DeepSeek | 19.1% |
| DeepSeek V4 Flash | #8 | DeepSeek | 1% |
| DeepSeek V4 Pro | #9 | DeepSeek | 0.4% |
One best score per model · higher is better
| Rank | Model | Provider | License | Evidence use | Score |
|---|---|---|---|---|---|
| #1 | Qwen3.7-Max qwen3.7-max | Alibaba Cloud | — | Reference only | 44.5% |
| #2 | DeepSeek V4 Pro (Max) deepseek-v4-pro-max | DeepSeek | open | Estimated reference | 38.3% |
| #3 | DeepSeek V4 Flash (Max) deepseek-v4-flash-max | DeepSeek | open | Estimated reference | 33% |
| #4 | ZAYA1-8B zaya1-8b | Zyphra | open | Estimated reference | 32.2% |
| #5 | DeepSeek V4 Pro (High) deepseek-v4-pro-high | DeepSeek | open | Estimated reference | 27.4% |
| #6 | Qwen3.7-Plus qwen3.7-plus | Alibaba Cloud | — | Reference only | 22.7% |
| #7 | DeepSeek V4 Flash (High) deepseek-v4-flash-high | DeepSeek | open | Estimated reference | 19.1% |
| #8 | DeepSeek V4 Flash deepseek-v4-flash | DeepSeek | open | Estimated reference | 1% |
| #9 | DeepSeek V4 Pro deepseek-v4-pro | DeepSeek | open | Reference only | 0.4% |
The top of this snapshot is led by Qwen3.7-Max at 44.5%; third place is 11.5 points behind. The top-9 spread is 44.1 points.
About Apex
A high-difficulty mathematical reasoning benchmark reported in DeepSeek-V4 model evaluations. Results stay tied to the exact model variant and evaluation system. Multiple systems for the same model use the best published score on this page; overall Lumina scoring uses the median of ranking-eligible rows.
Open benchmark source ↗FAQ
A high-difficulty mathematical reasoning benchmark reported in DeepSeek-V4 model evaluations.
Qwen3.7-Max by Alibaba Cloud currently leads with 44.5%.
9 models in the LuminaBench cohort have a qualifying score on this benchmark.
No. This benchmark is display-only and does not enter the overall Lumina composite.
Related
