B.AI’s daily token throughput reaches as high as 1.537 billion; MiniMax’s M3 continues to top the popularity chart.
TechFlow News, June 26: B.AI, an AI infrastructure platform, disclosed its latest core operational metrics, demonstrating robust underlying compute power support: the platform continues operating at ultra-high inference load, achieving a daily token throughput of 15.37 billion—maintaining this high level for multiple consecutive days and fully supporting large-scale AI applications and industrial-grade production workflows. API calls now account for 99.6% of total platform usage, with massive numbers of cutting-edge AI agents, complex automation pipelines, and mission-critical enterprise systems anchoring deeply into B.AI as their foundational infrastructure. Additionally, within the platform’s thriving model ecosystem, MiniMax’s M3 large language model consistently ranks first in call volume, showcasing exceptional inference efficiency and system stability even under extreme high-concurrency conditions.
B.AI is accelerating the global large language model ecosystem’s transition from technical validation to full-scale commercialization through solid underlying technological barriers.
