PolyReal: leaderboard
Metric: Overall F1 score of the key-point PRF framework (completeness and correctness against expert Key Points, LLM-judged) over the five PolyReal workflow modules, 545 questions, zero-shot, temperature 0; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 15 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | O3 | 0.6 |
| 2 | GPT-5 | 0.58 |
| 3 | Claude Sonnet 4.5 (Thinking) | 0.55 |
| 4 | Qwen 3 VL 235B A22B (Thinking) | 0.54 |
| 5 | Gemini 2.5 Pro | 0.54 |
| 6 | Grok 4 | 0.53 |
| 7 | Qwen 3 VL 32B (Thinking) | 0.53 |
| 8 | Gemini 2.5 Flash | 0.51 |
| 9 | DeepSeek R1 | 0.48 |
| 10 | Gemini 2.0 Flash (Thinking) | 0.47 |
| 11 | GPT-4o | 0.37 |
| 12 | Qwen 2.5 VL 72B Instruct | 0.34 |
| 13 | GPT-4o Mini | 0.3 |
Interactive version: theaggregate.ai/benchmark?slug=polyreal · How It Works · Data refreshed daily, snapshot 2026-10-07.