AI Decathlon - Open Design Tasks: leaderboard

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

Metric: Rubric-Graded Score (0-100 across 6 open design tasks). Source: octathlon.wendystars.com. Saturation forecast: Around February 2027. 25 models tracked. Updated 22 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1Claude Opus 5.5xHigh79.8#9
2Claude Fable 5.2xHigh79.8#2
3GPT-6.2 AstraxHigh79.1#4
4GPT-6.2 SolxHigh78.4#10
5Gemini 4 NeonxHigh76.1#3
6Claude Sonnet 5.5xHigh73.7#24
7GPT-6 AstraxHigh73.5#11
8Gemini 4 HydrogenxHigh72.6#16
9Gemini 4 CarbonHigh70.3#7
10GPT-6.1 SolxHigh70.2#15
11Claude Haiku 5.5xHigh61.6#75
12Kimi K3.1 ProMax58.1#35
13DeepSeek V4.1 FlashMax56.4#69
14GPT-6.1 TerraxHigh53.8#25
15hy4High46.7#44

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