MolViBench - Level 1 Remember and Understand: leaderboard
The Level 1 slice of MolViBench (2026), a molecular vibe coding benchmark: RDKit tasks on remembering and understanding: basic RDKit API recall, format conversion and single-step descriptor extraction (75 tasks). The score is Pass@1, the percent of tasks whose generated Python code reproduces the reference output, with each model generating once without execution feedback; higher is better.
Metric: Pass@1 (%) on the 75 Level 1 tasks (Remember and Understand), direct generation in one pass with no execution feedback, RDKit molecular coding tasks answered by generated Python code, official APIs, temperature 0, at most 4,096 output tokens; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 7 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Claude Opus 4.6 | 78.7 | #113 | |
| 2 | Gemini 3 Pro | 76 | #145 | |
| 3 | GPT-5.3 Codex | 76 | #124 | |
| 4 | GPT-5.2 Codex | 76 | #182 | |
| 5 | Kimi K2.5 | 74.7 | #266 | |
| 6 | DeepSeek V3.2 | Thinking | 69.3 | #419 |
| 7 | MiniMax-M2.5 | 64 | #600 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=molvibench-level-1-remember-and-understand · How It Works · Data refreshed daily, snapshot 2026-11-21.