GR-Ben (Thinking) - Inductive Reasoning: leaderboard

GR-Ben (Thinking) Inductive Reasoning scores LLM critics with slow thinking enabled on the GR-Ben (2026) subset built from the inductive-reasoning problems (MIRAGE), where the model must name the earliest erroneous step of a step-by-step solution or state that every step is correct. The score is F1 in percent, the harmonic mean of accuracy on flawed and on correct solutions, higher is better.

Metric: F1 (%) of first-error identification: the harmonic mean of the accuracy on flawed solutions (the earliest erroneous step named) and the accuracy on fully correct solutions (judged correct), on the 400 inductive-reasoning problems (MIRAGE) of GR-Ben (3,595 human-annotated solutions written by 15 LLMs); LLMs prompted as critics with slow thinking enabled (GPT-5.2 at xhigh and Gemini 3 Flash at high reasoning effort, 32,768 output tokens); higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 7 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 3 FlashHigh75.3#175
2GPT-5.2 (2025-12-11)xHigh73.7
3DeepSeek V3.2Thinking71.2#419
4Qwen 3 8B50.9#1873
5Qwen 3 32B49.1#972
6Qwen 3 4B45.6#2348
7Qwen 3 14B44.6#1269

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

Interactive version: theaggregate.ai/benchmark?slug=gr-ben-thinking-inductive-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-21.