List your company

Question Answering on GPQA Diamond (out-of-domain)

0.864AccuracyGemini
0.28680.436650.58650.73635Feb 4, 2026

Evaluation results

MethodLinks
Geminimode=Single-turn2026.020.864
OursOmode=Single-turn2026.020.781
GPTmode=Single-turn2026.020.701
OursImode=Single-turn2026.020.497
Qwenmode=Single-turn2026.020.309