benchgap
Calibration

FrontierMath v2 (Tier 4) → HMMT Feb 2026

HMMT Feb 2026 is estimated from FrontierMath v2 (Tier 4) with a linear curve fitted on 6 models measured on both: y = 0.3758·x + 0.8353, R² = 0.91, cross-validated error 1.5 pp. It is used for 33 estimates.

Estimated modelFrontierMath v2 (Tier 4)HMMT Feb 2026Source
Claude Haiku 5.546.3%100.0%estimated ± 1.5 pp, low confidence
Claude Opus 4.856.1%100.0%estimated ± 1.5 pp, low confidence
GPT-5.449.0%100.0%estimated ± 1.5 pp, low confidence
GPT-5.4 Pro58.5%100.0%estimated ± 1.5 pp, low confidence
GPT-5.6 Luna58.5%100.0%estimated ± 1.5 pp, low confidence
GPT-6 Luna56.1%100.0%estimated ± 1.5 pp, low confidence
Muse Spark 1.346.3%100.0%estimated ± 1.5 pp, low confidence
Qwen3.8 Max46.3%100.0%estimated ± 1.5 pp, low confidence
Gemini 3.7 Flash36.6%97.3%estimated ± 1.5 pp, low confidence
Claude Opus 4.731.7%95.4%estimated ± 1.5 pp, medium confidence
Grok 4.631.7%95.4%estimated ± 1.5 pp, medium confidence
GPT-5.231.7%95.4%estimated ± 1.5 pp, medium confidence
Claude Sonnet 529.3%94.5%estimated ± 1.5 pp, medium confidence
GLM-5.329.3%94.5%estimated ± 1.5 pp, medium confidence
Claude Opus 4.626.8%93.6%estimated ± 1.5 pp, medium confidence
DeepSeek V4.1 Flash 101726.8%93.6%estimated ± 1.5 pp, medium confidence
Gemini 3.1 Pro26.8%93.6%estimated ± 1.5 pp, medium confidence
Gemini 3.5 Flash26.8%93.6%estimated ± 1.5 pp, medium confidence
Grok 4.524.4%92.7%estimated ± 1.5 pp, medium confidence
Gemini 3.6 Flash22.0%91.8%estimated ± 1.5 pp, medium confidence
Gemini 3.8 Flash22.0%91.8%estimated ± 1.5 pp, medium confidence
Gemini 3 Flash17.1%89.9%estimated ± 1.5 pp, medium confidence
GLM-5.3-Flash17.1%89.9%estimated ± 1.5 pp, medium confidence
Grok 4.717.1%89.9%estimated ± 1.5 pp, medium confidence
Grok 4.314.6%89.0%estimated ± 1.5 pp, medium confidence
GPT-5.4 nano12.2%88.1%estimated ± 1.5 pp, medium confidence
Kimi K2.7 Code12.2%88.1%estimated ± 1.5 pp, medium confidence
GPT-5.4 mini9.8%87.2%estimated ± 1.5 pp, medium confidence
Inkling4.9%85.4%estimated ± 1.5 pp, medium confidence
o4-mini4.9%85.4%estimated ± 1.5 pp, medium confidence
Claude Sonnet 4.52.4%84.5%estimated ± 1.5 pp, low confidence
Gemini 2.5 Pro0.0%83.5%estimated ± 1.5 pp, low confidence
Gemini 3.5 Flash-Lite0.0%83.5%estimated ± 1.5 pp, low confidence

Fitted on

ModelFrontierMath v2 (Tier 4)HMMT Feb 2026
Claude Opus 4.54.9%85.3%
DeepSeek V4 Pro26.8%95.2%
GLM-5.229.3%92.5%
Inkling-Small17.1%90.2%
Kimi K2.625.6%92.7%
Qwen3.7 Max34.1%97.1%