MODEL PROFILE

LongCat 2.0

meituan · longcat-2.0

LongCat 2.0 is a sparse mixture-of-experts language model from Meituan, with 48B active parameters out of 1.6T total. It is suited for coding, repository-level changes, long-horizon problem solving, and agentic...

Text generationreasoningtools
Context window1,048,756 Token
Input → OutputText → Text
Model weightsOpen weights
Profile updated2026/11/8 00:54:17
OVERVIEW

Model overview

LongCat 2.0 is a sparse mixture-of-experts language model from Meituan, with 48B active parameters out of 1.6T total. It is suited for coding, repository-level changes, long-horizon problem solving, and agentic...

The summary below is compiled from sources by an AI model, not our own testing. Verify selection decisions against original sources and your workload.

Key features

  • Sparse mixture-of-experts architecture with 48B active parameters selected from 1.6T total parameters
  • Optimized for coding tasks, repository-level code changes, and long-horizon problem solving
  • Supports agentic workflows and tool-augmented reasoning

Limitations & considerations

Sources are insufficient to establish specific limitations. Verify task performance, latency, data policies, licensing and version lifecycle.

PRICING & COST

Source prices & cost analysis

Prices vary by channel, region and configuration. Channel quotes are listed independently and are not treated as a single official provider price.

Source / channelInputOutputBilling unitCollected at
OpenRouter · meituan/longcat-2.0 ↗USD 0.3USD 1.2Per million tokens2026/11/2 06:25:22

Some channels offer cache discounts. Check the source for cache read, write and additional fees.

Estimate monthly base token costs

—

Estimates exclude cache discounts, tiered pricing, multimodal surcharges, web search, taxes and additional reasoning output.

INDEPENDENT BENCHMARKS

Public benchmarks & run configuration

暂无已采集的可核验公开评分。模型未上榜不代表能力较低。

View full indexed rankings →

Frequently asked questions

What does '48B active parameters out of 1.6T total' mean?

It means the model uses a sparse MoE design where only 48 billion parameters are activated per forward pass, drawn from a total parameter pool of 1.6 trillion.

Is LongCat 2.0 open source?

The license is MIT, as stated in the provided data; however, no further details about source code availability or model weights distribution are given.

What types of tasks is LongCat 2.0 specifically designed for?

It is explicitly suited for coding, repository-level changes, long-horizon problem solving, and agentic applications.

Do open weights permit free commercial use?

Not necessarily. Check the version license, commercial conditions and usage policies. Self-hosting incurs compute and maintenance costs.

Are these real-time billing prices?

These are timestamped source snapshots, not live billing prices. Check the channel for current versions, regions, discounts and billing methods.

Sources & traceability

许可证:mit。模型介绍与第三方评测均以原来源的具体版本和方法为准。