List your company

Scientific Reasoning on GPQA D (Score)

85.4ScoreConfident Decoding
-3.41619.64242.765.758Jun 20, 2026

Evaluation results

MethodLinks
Confident DecodingModel=Qwen3.5-122B-A10B, Decoding Method=Confident2026.0685.4
Last Layer DecodingModel=Qwen3.5-122B-A10B, Decoding Method=Last Layer2026.0683.3
Confident DecodingModel=Qwen3.5-35B-A3B, Decoding Method=Confident2026.0682.8
Confident DecodingModel=Gemma-4-31B, Decoding Method=Confident2026.0680.8
Confident DecodingModel=gpt-oss-120B, Decoding Method=Confident2026.0680.8
Last Layer DecodingModel=Qwen3.5-27B, Decoding Method=Last Layer2026.0680
Confident DecodingModel=Qwen3.5-27B, Decoding Method=Confident2026.0680
Last Layer DecodingModel=Gemma-4-31B, Decoding Method=Last Layer2026.0676.8
Last Layer DecodingModel=Qwen3.5-35B-A3B, Decoding Method=Last Layer2026.0676.3
Last Layer DecodingModel=gpt-oss-120B, Decoding Method=Last Layer2026.0676.3
Last Layer DecodingModel=Qwen3.5-9B-Base, Decoding Method=Last Layer2026.0667.7
Confident DecodingModel=Qwen3.5-9B-Base, Decoding Method=Confident2026.0665.7
Last Layer DecodingModel=Qwen3.5-9B, Decoding Method=Last Layer2026.0664.6
Confident DecodingModel=Qwen3.5-9B, Decoding Method=Confident2026.0662.1
Confident DecodingModel=gpt-oss-20b, Decoding Method=Confident2026.0660.1
Last Layer DecodingModel=gpt-oss-20b, Decoding Method=Last Layer2026.0657.1
Last Layer DecodingModel=Qwen3.5-0.8B, Decoding Method=Last Layer2026.060
Confident DecodingModel=Qwen3.5-0.8B, Decoding Method=Confident2026.060