List your company

Scientific Instruction Following on SciRIFF

53.6Accuracyo3
46.63248.44150.2552.059Aug 26, 2025

Evaluation results

MethodLinks
o3Reasoning Effort=High2025.0853.61.8
Claude-Sonnet-4Reasoning Effort=Low2025.0853.5—
o4-miniReasoning Effort=High2025.0852.21.6
o3Reasoning Effort=Low2025.0851.8—
o3-miniReasoning Effort=High2025.0851.80.5
Gemini-2.5-ProReasoning Effort=Low2025.0851.6—
Gemini-2.5-ProReasoning Effort=High2025.0851.4-0.2
o3-miniReasoning Effort=Low2025.0851.3—
Claude-Sonnet-4Reasoning Effort=High2025.0850.9-2.6
o4-miniReasoning Effort=Low2025.0850.6—
GPT-5Reasoning Effort=High2025.0850.13.3
GPT-5Reasoning Effort=Low2025.0846.9—