B.AI October Ecosystem Monthly Report: Comprehensive Expansion of the Multimodal Flagship Matrix; New Additions to the Large Model L

B.AI, an AI model aggregation platform, reported registered users reaching 2,920,008 in October, approaching the 3 million mark, while hosting over 70 multimodal large language models accessible via API and Web Chat. The platform launched a new large model leaderboard based on real token consumption and API call frequency, with Chinese models sweeping the top eight spots in call rankings—DeepSeek-V4.1-Flash held the No. 1 position with 81.1% of weekly token usage (Oct 26–Nov 1), followed by Xiaomi's MiMo-V2.6-Flash as the fastest-growing model. On pricing, GPT-5 nano topped weighted and input price rankings, while Chinese models dominated output pricing, with MiniMax-M3.1-Flash at $0.30 per million tokens, GLM-5.3-Flash at $0.35, and Qwen3.8-Flash at $0.376. In intelligence rankings, Moonshot's Kimi K3.1 Pro scored 55.1 to enter the global top three behind Claude Opus 5.5 and Claude Sonnet 5.5, while Alibaba Cloud's Qwen3.8-Max ranked fifth at 53.4. The platform also added ByteDance's Seedance video generation series (up to 4K output), GPT-6.1 Terra, Claude Haiku 5.5, and MiniMax M3.1 Flash, and rolled out three rounds of discounts as low as 90% off. B.AI's Android App v1.3.0 now supports 11 languages, and its leaderboard data suggests intensifying competition among Chinese model providers for developer traffic in high-concurrency, cost-sensitive workloads.