URSA - Expert-2026 Solv-2: leaderboard
Metric: Solv-2 solved targets (%; share of the 100 URSA-expert-2026 novel, not-yet-synthesized drug-like targets with stock-terminated routes proposed for those targets whose every reaction also passes the ChemCensor precedent-based plausibility check (Solv-2), best of up to 10 routes per target, URSA-minor-1.1.0-U2 configuration with ChemCensor v1.2.0 and the U2 USPTO precedent database, shared stock of 255,365 building blocks; default reasoning effort; higher is better). Source: arxiv.org. Saturation forecast: Around January 2028. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 | 21 |
| 2 | Gemini 3.1 Pro (Preview) | 13 |
| 3 | Claude Opus 4.8 | 10 |
| 4 | Claude Opus 4.7 | 8 |
| 5 | Grok 4.1 | 7 |
| 6 | Claude Opus 4.6 | 5 |
| 7 | Kimi K2.5 | 5 |
| 8 | Claude Sonnet 4.6 | 3 |
| 9 | Claude Opus 4.5 | 3 |
| 10 | Qwen 3.5 397B A17B | 2 |
| 11 | GLM-5 | 2 |
| 12 | Claude Sonnet 4.5 | 1 |
| 13 | GPT-5.1 | 1 |
| 14 | GPT-5.4 | 0 |
| 15 | GPT-5.2 | 0 |
Interactive version: theaggregate.ai/benchmark?slug=ursa-expert-2026-solv-2 · How It Works · Data refreshed daily, snapshot 2026-09-29.