Models / LongCat-2.0
Meituan (LongCat)
LongCat-2.0
Specs as of
Meituan’s LongCat-2.0, released 30 Jun 2026: a 1.6T MoE with open weights, 1M context, and Meituan’s claim about training on a domestic-only cluster.
Specifications
- Lab
- Meituan (LongCat)
- Released
- 30 Jun 2026 (open-sourced at release)
- Parameters
- 1.6T total, about 48B active (MoE)
- Context window
- 1M tokens
- License
- Open-weight
- Pricing
- n/d
The benchmark claim
Meituan reports it as the first trillion-parameter model trained and inferenced entirely on a 50,000-card domestic cluster with no foreign accelerators, on over 30 trillion tokenslab-reported
A figure tagged lab-reported is the lab's own claim. A figure taggedindependent comes from a third-party leaderboard. We never present a vendor's number as neutral fact.
Sources
SCMP; Reuters; Geek Park; InfoQ
Covered in the daily: read the issue →
Every figure above traces to the named source on this page, and where it comes from our reporting, to the dated issue that carried it. Specs the lab did not disclose read "n/d" rather than carrying an estimate.