List your company

Question Answering on GPQA No-think (test)

47.8AccuracyQwen3-8B
34.17637.71341.2544.787Jan 11, 2026

Evaluation results

MethodLinks
Qwen3-8BTraining=RL, Mode=No-Think2026.0147.84,4359.863
Qwen3-4BTraining=RL, Mode=No-Think2026.0140.46,87117.196
Qwen3-4BTraining=RL, Mode=Mid-Think2026.0138.22,0864.79
Qwen3-8BTraining=No, Mode=Untrained2026.0137.51,3350.459
Qwen3-4BTraining=No, Mode=Untrained2026.0136.91,5600.59
Qwen3-8BTraining=RL, Mode=Mid-Think2026.0136.21,2931.968
Qwen3-4BTraining=RL, Mode=Think2026.0135.51,7890.252
Qwen3-8BTraining=RL, Mode=Think2026.0134.71,5672.179