Reconstruction (Research Idea Recovery) - Medicine: leaderboard
Metric: Match rate (%; the 78 medicine seed papers; share of a model's five hypotheses, proposed in one call from a published paper's pre-publication bibliography only (titles and abstracts under anonymous IDs, the paper itself and later literature withheld), that an LLM judge matches to the held-out paper's idea; mean over the six other evaluated models as leave-one-out judges). Source: arxiv.org. Saturation forecast: Around 2029. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.6 Pro Sol | 14.6 |
| 2 | Claude Opus 4.8 | 14.2 |
| 3 | Kimi K3 | 11.1 |
| 4 | Gemini 3.1 Pro (Preview) | 10.5 |
| 5 | GLM-5.2 | 10.1 |
| 6 | DeepSeek V4 Pro | 6.9 |
| 7 | Qwen 3.7 Max | 4.8 |
Interactive version: theaggregate.ai/benchmark?slug=reconstruction-research-idea-recovery-medicine · How It Works · Data refreshed daily, snapshot 2026-09-26.