List your company

Science Question Answering on ARC (Accuracy, Reusability, Verifiability)

95AccuracyDeepSeek-R1
80.4484.228891.78Feb 19, 2026

Evaluation results

MethodLinks
DeepSeek-R12026.02953263
Gemma32026.02933650
Llama3.12026.02834347
Phi4-Reasoning2026.02814370