Post: LongCat: Frontier reasoning models from Meituan

LongCat: Frontier reasoning models from Meituan

Congratulations to the Longcat team!

Being based in China, Meituan is a household name for me — I use it almost every week for food delivery and local services. So it’s honestly a little surprising, in a good way, to see a 1.6T MoE model from the same company open sourced under MIT. The contrast makes this launch particularly interesting.

I’m also exploring AI-assisted PM workflows and building my agent skills, so the combination of open weights, 1M context, and post-training for coding/agent workflows really stands out. If the team has a best practice for using LongCat-2.0 in Long Horizon agent workflows where the context includes product documentation, code, and task history.