Scientific Reasoning on GenomeBench
70.65AccuracyGPT-5-High*
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-5-High*Size (# Param.)=100B+, Reasoning Effort=high, Temperature=0.02025.09 | 70.65 | |
| Gemma-4B-Finetuned (NanoFlux-200)Size (# Param.)=4B, FLOPs=2.57×10^162025.09 | 61.3 | |
| Gemma-4B-Finetuned (Full Dataset)Size (# Param.)=4B, FLOPs=8.49×10^162025.09 | 57.6 | |
| Gemma-4BSize (# Param.)=4B2025.09 | 52.8 |