List your company

Instruction Following on Air-Bench

90ScoreConfident Decoding
49.4459.9770.581.03Jun 20, 2026

Evaluation results

MethodLinks
Confident DecodingModel=gpt-oss-120B, Decoding Method=Confident2026.0690
Confident DecodingModel=gpt-oss-20b, Decoding Method=Confident2026.0689
Last Layer DecodingModel=gpt-oss-120B, Decoding Method=Last Layer2026.0688
Last Layer DecodingModel=gpt-oss-20b, Decoding Method=Last Layer2026.0686
Last Layer DecodingModel=Qwen3.5-9B-Base, Decoding Method=Last Layer2026.0676
Confident DecodingModel=Qwen3.5-9B-Base, Decoding Method=Confident2026.0676
Confident DecodingModel=Gemma-4-31B, Decoding Method=Confident2026.0676
Last Layer DecodingModel=Gemma-4-31B, Decoding Method=Last Layer2026.0674
Last Layer DecodingModel=Qwen3.5-27B, Decoding Method=Last Layer2026.0667
Confident DecodingModel=Qwen3.5-122B-A10B, Decoding Method=Confident2026.0667
Last Layer DecodingModel=Qwen3.5-122B-A10B, Decoding Method=Last Layer2026.0666
Confident DecodingModel=Qwen3.5-27B, Decoding Method=Confident2026.0664
Confident DecodingModel=Qwen3.5-9B, Decoding Method=Confident2026.0656
Confident DecodingModel=Qwen3.5-35B-A3B, Decoding Method=Confident2026.0656
Confident DecodingModel=Qwen3.5-0.8B, Decoding Method=Confident2026.0654
Last Layer DecodingModel=Qwen3.5-9B, Decoding Method=Last Layer2026.0653
Last Layer DecodingModel=Qwen3.5-0.8B, Decoding Method=Last Layer2026.0652
Last Layer DecodingModel=Qwen3.5-35B-A3B, Decoding Method=Last Layer2026.0651