Mathematics · Benchmark profile
The most recent February edition of the Harvard-MIT Mathematics Tournament, featuring the latest challenging problems in competitive mathematics.
Data verified 27 Jul 2026 · Methodology 1.8.0
Benchmark score on Harvard-MIT Mathematics Tournament February 2025
GLM-5 leads at 97.5%, followed by Qwen3.6 Plus (96.7%) and Kimi K2.5 (95.4%).
7 ranked models · higher is better · labels show rank and score
| Model | Rank | Provider | Score |
|---|---|---|---|
| GLM-5 | #1 | Z.AI | 97.5% |
| Qwen3.6 Plus | #2 | Alibaba Cloud | 96.7% |
| Kimi K2.5 | #3 | Moonshot AI | 95.4% |
| Qwen3.5 397B | #4 | Alibaba Cloud | 94.8% |
| Qwen3.6 27B | #5 | Alibaba Cloud | 93.8% |
| Claude Opus 4.5 | #6 | Anthropic | 92.9% |
| Qwen3.6-35B-A3B | #7 | Alibaba Cloud | 90.7% |
One best score per model · higher is better
| Rank | Model | Provider | License | Evidence use | Score |
|---|---|---|---|---|---|
| #1 | GLM-5 glm-5 | Z.AI | open | Reference only | 97.5% |
| #2 | Qwen3.6 Plus qwen3-6-plus | Alibaba Cloud | closed | Reference only | 96.7% |
| #3 | Kimi K2.5 kimi-k2.5 | Moonshot AI | open | Reference only | 95.4% |
| #4 | Qwen3.5 397B qwen3.5-397b-a17b | Alibaba Cloud | open | Estimated reference | 94.8% |
| #5 | Qwen3.6 27B qwen3.6-27b | Alibaba Cloud | open | Estimated reference | 93.8% |
| #6 | Claude Opus 4.5 claude-opus-4-5 | Anthropic | closed | Reference only | 92.9% |
| #7 | Qwen3.6-35B-A3B qwen3-6-35b-a3b | Alibaba Cloud | open | Estimated reference | 90.7% |
The top of this snapshot is led by GLM-5 at 97.5%; third place is 2.1 points behind. The top-7 spread is 6.8 points.
About Harvard-MIT Mathematics Tournament February 2025
The most recent February edition of the Harvard-MIT Mathematics Tournament, featuring the latest challenging problems in competitive mathematics. 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
The most recent February edition of the Harvard-MIT Mathematics Tournament, featuring the latest challenging problems in competitive mathematics.
GLM-5 by Z.AI currently leads with 97.5%.
7 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
