List your company

Math Problem Solving on GSM8K (solve rate)

90.22Solve RateBifrost
0.249623.607346.96570.3227Feb 5, 2026

Evaluation results

MethodLinks
BifrostModel=Llama-3.1-8B2026.0290.22
ReflexionModel=Llama-3.1-8B2026.0287.41
RISE (5 attempts)Model=Llama-3.1-8B2026.0286.13
BifrostModel=Llama-3.2-3B2026.0281.41
ICLModel=Llama-3.1-8B2026.0280.78
RISE (5 attempts)Model=Llama-3.2-3B2026.0278.62
DoTModel=Llama-3.2-3B2026.0276.8
DoTModel=Llama-3.1-8B2026.0276.72
CoTModel=Llama-3.1-8B2026.0274.65
CoTModel=Llama-3.2-3B2026.0270.18
ICLModel=Llama-3.2-3B2026.0262.52
ReflexionModel=Llama-3.2-3B2026.0262.4
PaprikaModel=Llama-3.1-8B2026.0255.27
RISE (1 attempt)Model=Llama-3.1-8B2026.0248.29
BifrostModel=Llama-3.2-1B2026.0244.66
CoTModel=Llama-3.2-1B2026.0242.15
RISE (5 attempts)Model=Llama-3.2-1B2026.0240.71
PaprikaModel=Llama-3.2-3B2026.0240.26
BoTModel=Llama-3.1-8B2026.0239.73
RISE (1 attempt)Model=Llama-3.2-3B2026.0236.92
BoTModel=Llama-3.2-3B2026.0233.74
ICLModel=Llama-3.2-1B2026.0223.81
PaprikaModel=Llama-3.2-1B2026.0221.08
ReflexionModel=Llama-3.2-1B2026.0215.85
RISE (1 attempt)Model=Llama-3.2-1B2026.0214.48
BoTModel=Llama-3.2-1B2026.029.62
DoTModel=Llama-3.2-1B2026.023.71