ResearchBenchmarksPour on scientific experiment datasetFollow0.645Success RateCAPER0.336120.416310.49650.57669Feb 10, 2026Evaluation ResultsMethodMethodLinksSuccess RateCAPER2026.020.645MLM+RLBackbone=GPT-4oBackbone=GPT-4o2026.020.626MP+MLM+RLBackbone=GPT-4oBackbone=GPT-4o2026.020.624LLM+RLBackbone=GPT-4oBackbone=GPT-4o2026.020.546MLM+RLBackbone=LLaMA3.2, Rea...Backbone=LLaMA3.2, Reasoning Strategy=w/o CoT2026.020.447LLM+RLBackbone=Llama-3.1-8B,...Backbone=Llama-3.1-8B, Reasoning Strategy=w/o CoT2026.020.348