all the models — AI benchmark observatory
← Benchmarks

LAB-Bench Practical

LAB-Bench practical is a public multiple-choice evaluation of everyday wet-lab and research skills, covering protocol troubleshooting, sequence reasoning, and cloning scenarios. Distinct from labbench2.

id lab-bench-practical · max 1 · 1 models reported

#ModelScore

No scores for this benchmark yet.