LongCat-Flash-Thinking
MeituanOpen WeightMIT · Commercial OK
描述
LongCat-Flash-Thinking is Meituan's reasoning model built on the LongCat-Flash foundation with 560B total parameters (MoE, ~27B activated). It introduces a training pipeline specifically tuned for advanced reasoning, featuring Re-thinking Mode that delivers parallel reasoning paths for sophisticated decision-making. Achieves strong performance on mathematical reasoning, agentic tool use, and formal theorem proving benchmarks.
发布日期
2025-09-22
参数规模
560.0B
上下文长度
—
支持模态
text
能力雷达图
80
general
70
coding
90
reasoning
68
science估算
70
agents
0
multimodal
Science 在缺少专门科学评测时使用推理能力代理估算。
排行榜排名
基准测试分数 (LLM Stats)
Agents
BFCL-v3
74.4%自报
Biology
GPQA
81.5%自报
Code
LiveCodeBench
79.4%自报
SWE-Bench Verified
59.4%自报
Communication
Tau2 Telecom
83.1%自报
Tau2 Retail
71.5%自报
Tau2 Airline
67.5%自报
Finance
MMLU-Pro
82.6%自报
General
MMLU-Redux
89.3%自报
Math
MATH-500
99.2%自报
AIME 2024
93.3%自报
AIME 2025
90.6%自报
Reasoning
ZebraLogic
95.5%自报
ARC-AGI
50.3%自报
AA 评测指数
暂无 AA 评测数据
LLM Stats 分类评分
Language90
Math90
Biology80
Chemistry80
Finance80
General80
Healthcare80
Legal80
Physics80
Reasoning80
Structured Output70
Tool Calling70
Agents70
Code70
Communication70
Frontend Development60
Spatial Reasoning50
Vision50
定价
暂无定价数据
速度
暂无速度数据
可用提供商
(LS 内部计价单位)暂无提供商数据