PaperMind - Experimental Interpretation: leaderboard
PaperMind (2026) tests reasoning over real scientific papers from seven domains, with answers rated 1 to 5 by a GPT-4o judge against references taken from the papers. Here the model writes the analysis paragraph for an experimental figure or table, compared with the discussion in the paper itself; higher is better.
Metric: Judge score for writing the analysis paragraph for an experimental figure or table given the paper's introduction, the paper's own discussion paragraph as reference; rated by a GPT-4o judge on a 5-point scale (1-5) against the reference, averaged over the seven scientific domains; higher is better. Source: arxiv.org. Saturation forecast: Around 2029. 7 models tracked. Updated 14 Nov 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 2.25 | #276 |
| 2 | Claude 3.5 Sonnet | 2.13 | #752 |
| 3 | GPT-4o Mini | 1.92 | #1667 |
| 4 | Qwen 3 VL 4B Instruct | 1.9 | #1332 |
| 5 | Claude 3 Haiku | 1.82 | #2663 |
| 6 | Gemma 3 4B (IT) | 1.73 | #4139 |
| 7 | Phi-3.5-Vision Instruct | 1.61 | #6268 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=papermind-experimental-interpretation · How It Works · Data refreshed daily, snapshot 2026-11-21.