List your company

Mathematical Reasoning on Comp. Math (val)

62.7AccuracyBase
44.39649.14853.958.652Apr 10, 2026

Evaluation results

MethodLinks
BaseBackbone=Qwen3-32B, Generations=642026.0462.73.6734.7
ControlBackbone=Qwen3-32B, Generations=642026.0458.54.5148.5
MEMENTOBackbone=Qwen3-32B, Generations=642026.0456.21.7415.7
BaseBackbone=Phi-4-r (14B), Generations=642026.0455.13.0635.9
BaseBackbone=Qwen3-8B, Generations=642026.0454.32.7130.9
BaseBackbone=Olmo 3 (7B), Generations=82026.0452.74.2160.1
ControlBackbone=Phi-4-r (14B), Generations=642026.0451.43.4836.7
Mem. + RLBackbone=Qwen3-8B, Generations=642026.0449.41.4816.4
ControlBackbone=Qwen3-8B, Generations=642026.0449.22.8233.1
MEMENTOBackbone=Phi-4-r (14B), Generations=642026.0448.71.2513.1
ControlBackbone=Olmo 3 (7B), Generations=82026.0448.33.7846
MEMENTOBackbone=Olmo 3 (7B), Generations=82026.0448.13.4343.6
MEMENTOBackbone=Qwen3-8B, Generations=642026.0445.11.0810.7