PolyBench - Advanced Property Reasoning: leaderboard
Metric: LLM-Judge Score (1-7 Likert; GPT-OSS-20B judge, open-ended answers). Source: arxiv.org. Saturation forecast: Around June 2027. 21 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.4 | 5.68 |
| 2 | Phi-4 | 5.19 |
| 3 | Claude Sonnet 4.6 | 4.88 |
| 4 | GPT-4o | 4.73 |
Interactive version: theaggregate.ai/benchmark?slug=polybench-advanced-property-reasoning · How It Works · Data refreshed daily, snapshot 2026-09-25.