List your company

Mathematical Reasoning on AIME 2024 (avg@32)

81.35avg@32Qwen3-Next-80B-A3B-Instruct
14.509231.862149.21566.5679Jan 29, 2026Jan 31, 2026Feb 3, 2026Feb 6, 2026Feb 9, 2026Feb 12, 2026Feb 15, 2026

Evaluation results

MethodLinks
Qwen3-Next-80B-A3B-InstructArchitecture=MoE, # Total Params=80B, # Activated Params=3B2026.0181.35
LongCat-Flash-LiteArchitecture=MoE + NE, # Total Params=68.5B, # Activated Params=2.9B~4.5B2026.0172.19
Kimi-Linear-48B-A3BArchitecture=MoE, # Total Params=48B, # Activated Params=3B2026.0170.52
Gemini 2.5 Flash-Lite2026.0163.33
PMD-MEANBackbone=Qwen3-30B-A3B-Base, tau=0.1, Staleness=162026.0250.83
PMD-MEANBackbone=Qwen3-30B-A3B-Base, tau=0.01, Staleness=162026.0250
RLBitwidth=BF162026.0240.73
QuRL w/ UAQBitwidth=INT82026.0240.52
QuRL w/o UAQBitwidth=INT82026.0239.06
FlashRLBitwidth=INT82026.0236.77
GRPOBackbone=Qwen3-30B-A3B-Base, tau=null, Staleness=162026.0236.56
RLBitwidth=INT82026.0233.95
BaseBitwidth=BF162026.0228.54
PMD-MEANBackbone=Qwen2.5-7B, tau=0.02, Staleness=162026.0222.5
PMD-MEANBackbone=Qwen2.5-7B, tau=0.005, Staleness=162026.0219.69
On-policyBackbone=Qwen2.5-7B, tau=null, Staleness=12026.0218.65
PMD-MEANBackbone=Qwen2.5-7B, tau=0.01, Staleness=162026.0217.6
GRPOBackbone=Qwen2.5-7B, tau=null, Staleness=162026.0217.08