GLM 5.3: Tops B.AI Leaderboard

Justin Sun announced GLM 5.3 has seized the #1 ranking on B.AI with cumulative token throughput exceeding 2.41 trillion tokens. The GLM-5.3-Flash variant runs 320B total parameters and 18B active ones under a hybrid sparse-linear attention design that supports a 1M-token context window, driving fast inference and strong cost efficiency for API calls, coding tasks and long-document agents. The model remains 100% free on the platform, accelerating large language model adoption and multimodal AI capabilities across high-volume users.