List your company

Multi-turn Dialogue on Spec-Bench Multi.

3.22CRSpecBound
1.54561.98032.4152.8497Apr 14, 2026

Evaluation results

MethodLinks
SpecBoundBase Model=Vicuna-13B2026.043.221.69
SpecBoundBase Model=Vicuna-7B2026.043.051.65
SpecBoundBase Model=Codellama-7B-Instruct2026.042.741.66
SpecBoundBase Model=Codellama-13B-Instruct2026.042.691.85
KangarooBase Model=Vicuna-13B2026.042.441.66
SPACEBase Model=Vicuna-7B2026.042.311.71
KangarooBase Model=Vicuna-7B2026.042.221.68
AdaDecodeBase Model=Codellama-13B-Instruct2026.042.21.52
RESTBase Model=Codellama-7B-Instruct2026.042.071.67
RESTBase Model=Vicuna-7B2026.042.041.65
AdaDecodeBase Model=Codellama-7B-Instruct2026.0421.4
RESTBase Model=Vicuna-13B2026.041.941.5
RESTBase Model=Codellama-13B-Instruct2026.041.911.48
MedusaBase Model=Codellama-13B-Instruct2026.041.892.09
MedusaBase Model=Vicuna-13B2026.041.862.07
LookaheadBase Model=Codellama-7B-Instruct2026.041.781.51
LookaheadBase Model=Vicuna-7B2026.041.741.49
MedusaBase Model=Vicuna-7B2026.041.721.95
MedusaBase Model=Codellama-7B-Instruct2026.041.681.92
LookaheadBase Model=Vicuna-13B2026.041.641.41
LookaheadBase Model=Codellama-13B-Instruct2026.041.611.39