ChemRxn-Bench Judge - RxnMiner Reaction F1 Strict: leaderboard
AppendixTable7 secondary judge: atom-level validity/semantic conditions separate from main rule-based exact-molecule scoring. RP judge precision,RR rule recall,RF1 per-figure judge macroF1; parentheses stereochemistry-blind. Weighted Overall judge composite. Same603figure manifest and zero-credit failure policy; source method-specific results.
Metric: Judge score (%). Source: arxiv.org. Saturation forecast: Around February 2027. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5 | 86.37 | #16 |
| 2 | Kimi K3 | 78.61 | #35 |
| 3 | DeepSeek V4.1 Flash | 73.51 | #37 |
| 4 | GPT-5.6 Sol | 63.66 | #22 |
| 5 | Seed 2.1 Pro | 52.38 | #73 |
| 6 | GLM-5.3 Flash | 29.33 | #77 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=chemrxn-bench-judge-rxnminer-reaction-f1-strict · How It Works · Data refreshed daily, snapshot 2026-11-05.