AI Decathlon - Math Reasoning: leaderboard

AI Decathlon Timed Challenge (Math Reasoning) evaluating coding agents on 60-100 minute engineering, math, spatial, and interactive design tasks.

Metric: Math Reasoning Score (0-100). Source: octathlon.wendystars.com. Saturation forecast: Around February 2027. 25 models tracked. Updated 22 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1GPT-6.2 AstraxHigh100#4
2GPT-6.2 SolxHigh100#10
3GPT-6 AstraxHigh75#11
4GPT-6.1 SolxHigh69.3#15
5Gemini 4 NeonxHigh68.8#3
6Claude Opus 5.5xHigh58.3#9
7Claude Fable 5.2xHigh58.3#2
8Gemini 4 CarbonHigh54.2#7
9GPT-6.1 TerraxHigh48.2#25
10Gemini 4 HydrogenxHigh47.9#16
11Claude Sonnet 5.5xHigh41.4#24
12GPT-5.6 SolxHigh40.5#36
13Claude Haiku 5.5xHigh26.2#75
14Kimi K3.1 ProMax19.3#35
15GPT-6 LunaxHigh18.5#108

Interactive version: theaggregate.ai/benchmark?slug=ai-decathlon-math-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-22.