Inference

Harvard-MIT Mathematics Tournament February 2026

A February 2026 HMMT slice used in newer frontier-model math comparisons.

  • In index
  • Math
ModelsModels with a result.23
Top resultBest result on this test.97.1%Qwen3.7 Max
Top-3 spreadPoints from first to third.2.3 pts
YearYear of release.N/A

Result and price

020406080100$0.03$0.1$0.3$1$3$10$30RESULTOUTPUT PRICE PER 1M TOKENS · LOG SCALE

Results

23 results
#ModelResultThe score from the source.
01Qwen3.7 Max97.1%
02DeepSeek V4 Pro 081395.2%
03DeepSeek V4 Flash 073194.8%
04Solar Open 293.9%
05Qwen3.7 Plus92.9%
06Kimi K2.692.7%
07GLM-5.292.5%
08Inkling-Small90.2%
09Qwen3.5 397B87.9%
10Qwen3.6 Plus87.8%
11Kimi K2.587.1%
12Ling 3.0 Flash87.0%
13GLM-586.4%
14Claude Opus 4.585.3%
15MAI-Thinking-184.9%
16Qwen3.6-27B84.3%
17Qwen3.6-35B-A3B83.6%
18GLM-5.182.6%
19K-EXAONE 2.078.4%
20ZAYA1-8B71.6%
21MiniCPM5-2B63.8%
22LongCat-Flash-Lite-Sparse40.5%
23MiniCPM5-1B25.8%