ByteBrief
We're a portrait publication through and through. Turn your phone back and your briefing picks up right where you left it.
(We tried widescreen once. It wasn't us.)

Meituan released LongCat-2.0, an open-source Mixture-of-Experts language model with 1.6 trillion total parameters and a 1-million-token context window. The model was trained and served entirely on over 50,000 domestic AI ASIC superpods, avoiding Nvidia hardware. Meituan reports it edges GPT-5.5 on SWE-bench Pro and matches Google's Gemini 3.1 Pro in software engineering benchmarks.
Tracked by ByteBrief