Harvard-MIT Mathematics Tournament February 2026 (HMMT Feb 2026)
A February 2026 HMMT slice used in newer frontier-model math comparisons.
Top models on HMMT Feb 2026 — October 29, 2026
As of October 29, 2026, Qwen3.7 Max leads the HMMT Feb 2026 leaderboard with 97.1% , followed by DeepSeek V4 Pro 0813 (95.2%) and DeepSeek V4 Flash 0731 (94.8%).
Qwen3.7 Max
Alibaba
DeepSeek V4 Pro 0813
DeepSeek
DeepSeek V4 Flash 0731
DeepSeek
29 modelsMathematics25% of category scoreCurrentUpdated October 29, 2026
Leaderboard (29 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | Qwen3.7 MaxAlibaba | 97.1% | Not reported | Closed |
| 2 | DeepSeek V4 Pro 0813DeepSeek | 95.2% | Not reported | Open |
| 3 | DeepSeek V4 Flash 0731DeepSeek | 94.8% | Not reported | Open |
| 4 | DeepSeek V4 Pro (High)DeepSeek | 94.0% | Not reported | Open |
| 5 | Solar Open 2Upstage | 93.9% | Not reported | Open |
| 6 | Qwen3.7 PlusAlibaba | 92.9% | Not reported | Closed |
| 7 | Kimi K2.6Moonshot AI | 92.7% | Not reported | Open |
| 8 | GLM-5.2Z.AI | 92.5% | Not reported | Open |
| 9 | DeepSeek V4 Flash (High)DeepSeek | 91.9% | Not reported | Open |
| 10 | Inkling-SmallThinking Machines Lab | 90.2% | Not reported | Open |
| 11 | Qwen3.5 397BAlibaba | 87.9% | Not reported | Open |
| 12 | Qwen3.6 PlusAlibaba | 87.8% | Not reported | Closed |
| 13 | Kimi K2.5Moonshot AI | 87.1% | Not reported | Open |
| 14 | Ling 3.0 FlashInclusionAI | 87.0% | Not reported | Open |
| 15 | GLM-5Z.AI | 86.4% | Not reported | Open |
| 16 | Claude Opus 4.5Anthropic | 85.3% | Not reported | Closed |
| 17 | MAI-Thinking-1Microsoft | 84.9% | Not reported | Closed |
| 18 | Qwen3.6-27BAlibaba | 84.3% | Not reported | Open |
| 19 | Qwen3.6-35B-A3BAlibaba | 83.6% | Not reported | Open |
| 20 | GLM-5.1Z.AI | 82.6% | Not reported | Open |
| 21 | Spark-X2.6-4BSparkLLM | 81.3% | Not reported | Open |
| 22 | K-EXAONE 2.0LG AI Research | 78.4% | Not reported | Open |
| 23 | MiniMax M2.7MiniMax | 72.7% | Not reported | Open |
| 24 | ZAYA1-8BZyphra | 71.6% | Not reported | Open |
| 25 | MiniCPM5-2BOpenBMB | 63.8% | Not reported | Open |
| 26 | DeepSeek V4 FlashDeepSeek | 40.8% | Not reported | Open |
| 27 | LongCat-Flash-Lite-SparseMeituan | 40.5% | Not reported | Open |
| 28 | DeepSeek V4 ProDeepSeek | 31.7% | Not reported | Open |
| 29 | MiniCPM5-1BOpenBMB | 25.8% | Not reported | Open |
According to BenchLM.ai, Qwen3.7 Max leads the HMMT Feb 2026 benchmark with a score of 97.1%, followed by DeepSeek V4 Pro 0813 (95.2%) and DeepSeek V4 Flash 0731 (94.8%). The top models are clustered within 2.3 points, suggesting this benchmark is nearing saturation for frontier models.
29 models have been evaluated on HMMT Feb 2026. The benchmark falls in the Mathematics category. The Mathematics leaderboard ranks models by a weighted category score, and HMMT Feb 2026 contributes 25% of it. BenchAlign v5.8 also uses it in the overall ranking.
About HMMT Feb 2026
Year
2026
Tasks
Competition math problems
Format
Contest mathematics
Difficulty
Olympiad-style mathematics
HMMT February 2026 matters because small score deltas at the frontier often depend on which contest set is used. BenchLM keeps this newer slice distinct from older HMMT summary rows.
Freshness and provenance
Version
HMMT Feb 2026 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does HMMT Feb 2026 measure?
A February 2026 HMMT slice used in newer frontier-model math comparisons.
Which model scores highest on HMMT Feb 2026?
Qwen3.7 Max by Alibaba currently leads with a score of 97.1% on HMMT Feb 2026.
How many models are evaluated on HMMT Feb 2026?
29 AI models have published results on HMMT Feb 2026 in the BenchLM catalog.
Compare top models on HMMT Feb 2026
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 6,100+ subscribers.
One email each week. Unsubscribe anytime.