Meituan's LongCat-2.0-Preview, disguised as Owl Alpha on OpenRouter, processes 10.1T monthly tokens and ranks first globally, all without NVIDIA chips.

Chinese AI models now process over 18 trillion tokens weekly on OpenRouter versus 5.5 trillion for US models, marking a dramatic reversal from 2025's

Meituan's LongCat-2.0-Preview, disguised as Owl Alpha on OpenRouter, processes 10.1T monthly tokens and ranks first globally, all without NVIDIA chips.

LongCat-2.0 boasts 1.6 trillion parameters and a million-token context window, on par with DeepSeek’s latest flagship model.

LongCat-2.0 boasts 1.6 trillion parameters and a million-token context window, on par with DeepSeek’s latest flagship model.

Everyone argues about which AI model is "best" using benchmarks. I wanted to know something simpler:...

By registering the LongCat-2.0 repository under the open-source MIT License, Meituan positions the architecture with maximum legal flexibility for enterprise integration.

China achieves a breakthrough in AI development as Meituan unveils LongCat-2.0, the nation's largest model trained entirely on domestically-produced chips.

China’s Meituan open-sources massive LongCat-2.0 AI model, saying it was trained on domestic chips - SiliconANGLE