List your company

Long Reasoning on AIME 24 (pass@k, avg@k)

86.67Pass@kFull
45.0755.8766.6777.47May 19, 2025

Evaluation results

MethodLinks
FullBackbone=DeepSeek-R1-Qwen-14B, k=8, KV cache method=Full2025.0586.6766.25
FullBackbone=DeepSeek-R1-Qwen-7B, k=8, KV cache method=Full2025.0583.3356.66
FreeKVBackbone=DeepSeek-R1-Qwen-7B, k=8, KV cache method=FreeKV2025.0583.3352.92
QuestBackbone=DeepSeek-R1-Qwen-14B, k=8, KV cache method=Quest2025.0583.3358.33
ShadowKVBackbone=DeepSeek-R1-Qwen-14B, k=8, KV cache method=ShadowKV2025.0583.3357.25
FreeKVBackbone=DeepSeek-R1-Qwen-14B, k=8, KV cache method=FreeKV2025.0583.3364.17
FullBackbone=DeepSeek-R1-Llama-8B, k=8, KV cache method=Full2025.058047.08
ArkValeBackbone=DeepSeek-R1-Llama-8B, k=8, KV cache method=ArkVale2025.058046.67
FreeKVBackbone=DeepSeek-R1-Llama-8B, k=8, KV cache method=FreeKV2025.0576.6747.5
QuestBackbone=DeepSeek-R1-Qwen-7B, k=8, KV cache method=Quest2025.0576.6747.5
ArkValeBackbone=DeepSeek-R1-Qwen-14B, k=8, KV cache method=ArkVale2025.0576.6761.25
InfiniGenBackbone=DeepSeek-R1-Qwen-14B, k=8, KV cache method=InfiniGen2025.0576.6760
QuestBackbone=DeepSeek-R1-Llama-8B, k=8, KV cache method=Quest2025.0573.3344.17
RaaSBackbone=DeepSeek-R1-Qwen-7B, k=8, KV cache method=RaaS2025.0573.3342.92
ArkValeBackbone=DeepSeek-R1-Qwen-7B, k=8, KV cache method=ArkVale2025.0573.3347.92
ShadowKVBackbone=DeepSeek-R1-Qwen-7B, k=8, KV cache method=ShadowKV2025.0573.3343.75
RaaSBackbone=DeepSeek-R1-Qwen-14B, k=8, KV cache method=RaaS2025.0573.3348.75
InfiniGenBackbone=DeepSeek-R1-Llama-8B, k=8, KV cache method=InfiniGen2025.057045.83
RaaSBackbone=DeepSeek-R1-Llama-8B, k=8, KV cache method=RaaS2025.0566.6736.25
RazorBackbone=DeepSeek-R1-Qwen-7B, k=8, KV cache method=Razor2025.0565.3335.42
ShadowKVBackbone=DeepSeek-R1-Llama-8B, k=8, KV cache method=ShadowKV2025.0563.3336.5
InfiniGenBackbone=DeepSeek-R1-Qwen-7B, k=8, KV cache method=InfiniGen2025.0563.3343.34
RazorBackbone=DeepSeek-R1-Llama-8B, k=8, KV cache method=Razor2025.0546.6730
RazorBackbone=DeepSeek-R1-Qwen-14B, k=8, KV cache method=Razor2025.0546.6732.5