General365 - Recursive & Backtracking: leaderboard

General365 (Meituan LongCat team, 2026) poses 1,460 general-reasoning problems (365 hand-written seeds and their expansions) that need little expert knowledge; numeric answers are checked with math-verify and the others graded by GPT-4.1. This board covers the problems labelled Recursive & Backtracking (a problem can carry several labels). Accuracy in percent; higher is better.

Metric: Accuracy (%) on the problems labelled Recursive & Backtracking (a problem may carry several challenge labels), out of 1,460 problems (365 hand-written seeds and 1,095 expansions); numeric answers checked with math-verify, choice and text answers graded by GPT-4.1; temperature 1.0 for reasoning and 0.7 for chat models, highest available reasoning effort; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 25 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 3 Flash69.3#175
2GLM-4.7Thinking68.8#380
3GLM-5Thinking68.2#269
4Qwen 3.5 397B A17BThinking67.9#264
5Gemini 3 Pro66.8#145
6Qwen 3 MaxThinking66.8#415
7DeepSeek V3.2 Speciale65.8#273
8Kimi K2.5Thinking65.2#266
9GPT-5High65.2#178
10DeepSeek V3.2Thinking64.9#419
11GPT-5.1High64.4#240
12GLM-4.6Thinking64.1#498
13O4 MiniHigh62.8#368
14Kimi K2 (Thinking)62.5
15Grok 4.1 FastReasoning62.5#427

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

Interactive version: theaggregate.ai/benchmark?slug=general365-recursive-backtracking · How It Works · Data refreshed daily, snapshot 2026-11-21.