CAE-Bench RC7 - Fixed Problems: leaderboard

Metric: Score (0-100), 24 tasks. Source: arxiv.org. Saturation forecast: Estimated already saturated. 12 models tracked. Updated 18 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1GPT-6 AstraMedium100#7
2GPT-6.1 SolMedium100#12
3GPT-5.6 LunaxHigh100#129
4GPT-6 SolMedium100#30
5Gemini 3.8 FlashMedium98.55#39
6GPT-5.6 SolMedium97.22#33
7Gemini 4 ArgonMedium95.83#14
8GPT-6 LunaMax91.67#102
9DeepSeek V4.1 FlashHigh91.67#67
10Kimi K2.7 CodeThinking76.29#195
11GLM-5.3 FlashLow73.55#134
12Qwen 3.8 FlashLow70.83#127

No result here: #1 Claude Fable 5.2, #2 GPT-6.2 Astra, #3 Gemini 4 Carbon.

Interactive version: theaggregate.ai/benchmark?slug=cae-bench-rc7-fixed-problems · How It Works · Data refreshed daily, snapshot 2026-11-18.