GR-Ben - Inductive Reasoning: leaderboard
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); process reward models by their step predictions (scalar-score PRMs thresholded on a 40-item dev split held out of their scoring), LLMs prompted as critics with thinking disabled (mean of three runs for open models, one run for API models); higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 22 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Flash (Minimal) | 74.3 |
| 2 | DeepSeek V3.2 (Non-reasoning) | 67.5 |
| 3 | Kimi K2 0905 | 66.5 |
| 4 | GPT-5.2 (Non-reasoning) | 54.5 |
| 5 | Qwen 3 8B (Non-reasoning) | 37 |
| 6 | Gemma 3 27B (IT) | 34.9 |
| 7 | Qwen 3 14B (Non-reasoning) | 32.2 |
| 8 | Qwen 3 32B (Non-reasoning) | 32.1 |
| 9 | Qwen 3 4B (Non-reasoning) | 31.1 |
| 10 | Llama 3.3 70B Instruct | 28.8 |
| 11 | Gemma 3 12B (IT) | 23.1 |
Interactive version: theaggregate.ai/benchmark?slug=gr-ben-inductive-reasoning · How It Works · Data refreshed daily, snapshot 2026-10-07.