List your company

Scientific Reasoning on GPQA Diamond (Samples Required & Average Improvement)

8Samples RequiredBlend-ASC
4.2429.625580.38Nov 15, 2025

Evaluation results

MethodLinks
Blend-ASCSC@n=64, Model=Qwen-32B2025.1185.33
Adaptive SCSC@n=64, Model=Qwen-32B2025.11134.36
Fixed-AllocationSC@n=64, Model=Qwen-32B2025.11154.09
Blend-ASCSC@n=128, Model=Qwen-32B2025.11254.29
Blend-ASCSC@n=64, Model=Llama-3B2025.11265.33
Fixed-AllocationSC@n=64, Model=Llama-3B2025.11274.09
Adaptive SCSC@n=128, Model=Qwen-32B2025.11292.13
Adaptive SCSC@n=64, Model=Llama-3B2025.11334.36
Fixed-AllocationSC@n=128, Model=Qwen-32B2025.11373.4
Fixed-AllocationSC@n=128, Model=Llama-3B2025.11773.4
Blend-ASCSC@n=128, Model=Llama-3B2025.11804.29
Adaptive SCSC@n=128, Model=Llama-3B2025.111022.13