Chemistry Question Answering on ChemBench Chemical Preference
93.1Success RateMulti-Agentic Workflow for Science
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Multi-Agentic Workflow for ScienceUnderlying LLM=Claude Sonnet 4 model2026.01 | 93.1 | 56 | -4 | 5 |
Loading the SOTA2 catalog…