PolyReal - Performance and Application Exploration: leaderboard
PolyReal (2026) evaluates multimodal models on 545 expert-written polymer science tasks covering five stages of real experimental work. Answers are scored against expert key points by an LLM judge and summarized as F1 from 0 to 1. This board is F1 on the module about judging structure-property relations and suitability for an application; higher is better.
Metric: F1 score on the Performance and Application Exploration module of PolyReal (structure-property reasoning and material suitability); higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 15 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | GPT-5 | 0.623 | #178 | |
| 2 | O3 | 0.609 | #237 | |
| 3 | Gemini 2.5 Pro | 0.555 | #276 | |
| 4 | Grok 4 | 0.54 | #319 | |
| 5 | Claude Sonnet 4.5 | Thinking | 0.539 | #261 |
| 6 | Qwen 3 VL 32B (Thinking) | 0.538 | #573 | |
| 7 | Qwen 3 VL 235B A22B (Thinking) | 0.518 | #459 | |
| 8 | Gemini 2.5 Flash | 0.495 | #463 | |
| 9 | Intern-S1 | 0.476 | #624 | |
| 10 | Gemini 2.0 Flash (Thinking) | 0.468 | ||
| 11 | DeepSeek R1 | 0.427 | #482 | |
| 12 | Qwen 2.5 VL 72B Instruct | 0.402 | #841 | |
| 13 | GPT-4o | 0.37 | #751 | |
| 14 | GPT-4o Mini | 0.351 | #1667 | |
| 15 | DeepSeek-VL2 | 0.122 | #2927 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=polyreal-performance-and-application-exploration · How It Works · Data refreshed daily, snapshot 2026-11-21.