LegendBench - Reasoning: leaderboard
Legend-conditioned reasoning in LegendBench (Zhang et al., 2026): questions that require inference over legend-bound values. Raw accuracy in percent; higher is better.
Metric: Reasoning accuracy (%; raw accuracy on legend-conditioned questions that require inference; LegendBench snapshot of 26,689 legend-centric chart questions generated from 900 base charts with counterfactual legend variants; the same prompt template and hyperparameters for every model). Source: arxiv.org. Saturation forecast: Around February 2027. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.6 Sol | 74.16 | #29 |
| 2 | Qwen 3.5 27B | 56.79 | #358 |
| 3 | Gemini 2.5 Flash | 53.71 | #452 |
| 4 | GLM-4.6V | 53.41 | #698 |
| 5 | GPT-4.1 Mini | 50.38 | #775 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=legendbench-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-12.