PolyBench - Property Comparison & Ranking - Multiple-Choice Accuracy: leaderboard
Multiple-choice questions from PolyBench (Mohanty et al., Stony Brook University, 2026) that compare held-out polymers on measured properties or pick candidates that satisfy several structural and property constraints. The score is accuracy in percent; higher is better. Models suffixed PolyBench are the authors' fine-tunes on the benchmark's training split.
Metric: Accuracy (%; multiple-choice comparison and constrained screening). Source: arxiv.org. Saturation forecast: Around September 2027. 21 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.4 | 70 | #121 |
| 2 | GPT-4o | 69 | #741 |
| 3 | Claude Sonnet 4.6 | 66 | #158 |
| 4 | Phi-4 | 61 | #1920 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=polybench-property-comparison-ranking-multiple-choice-accuracy · How It Works · Data refreshed daily, snapshot 2026-11-12.