SuperCLUE General (November 2026) - Mathematical and Scientific Reasoning: leaderboard

SuperCLUE’s November 2026 general evaluation. The native 数理推理 column jointly covers mathematical and scientific reasoning; it is not either earlier separate subject column. Source workbook: https://www.superclueai.com/data/generalboard/2026%E5%B9%B411%E6%9C%88.xlsx. The publisher-reported score and exact native model/run labels are retained; higher is better. This question-set edition remains separate from earlier editions.

Metric: Score. Source: www.superclueai.com. Saturation forecast: Around January 2028. 19 models tracked. Updated 14 Nov 2026.

Top models

#ModelScoreOverall rank
1GPT-6.1 Sol76.32#10
2Darwin-180B-RSI-R373.68
3Gemini 4 Argon71.05#12
4DeepSeek V4.1 Flash (1017)71.05#41
5Grok 4.768.42#47
6Qwen 3.8 Max (0902)68.42#49
7Spark X2.668.42#170
8GLM-5.365.79#68
9MiniMax-M3.1-Flash65.79#109
10MiniMax-M363.16#272
11Kimi K3.1 Pro63.16#29
12Doubao-Seed-2.1-Pro-26091560.53
13GLM-5.3 Flash57.89#125
14MiMo-V2.6-Flash57.89#155
15Step 5 Preview57.89#110

No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.

Interactive version: theaggregate.ai/benchmark?slug=superclue-general-november-2026-mathematical-and-scientific-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-16.