List your company

Scientific Reasoning on GenomeBench

70.65AccuracyGPT-5-High*
52.08656.905561.72566.5445Sep 27, 2025

Evaluation results

MethodLinks
GPT-5-High*Size (# Param.)=100B+, Reasoning Effort=high, Temperature=0.02025.0970.65
Gemma-4B-Finetuned (NanoFlux-200)Size (# Param.)=4B, FLOPs=2.57×10^162025.0961.3
Gemma-4B-Finetuned (Full Dataset)Size (# Param.)=4B, FLOPs=8.49×10^162025.0957.6
Gemma-4BSize (# Param.)=4B2025.0952.8