B.AI Daily Token Throughput Surpasses 1.1 Trillion, AI Infrastructure Scaling Effects Accelerate

Recently, AI infrastructure platform B.AI surpassed 1.1 trillion tokens in single-day throughput for the first time. This marks not only a significant growth milestone in the nearly five months since the platform's launch, but also signals that B.AI has entered the ranks of trillion-level AI call infrastructure. As global developers continue to release demand for multi-model calls and Agent applications, B.AI is extending its role from a unified model gateway further into underlying layers such as model routing, call metering, and payment settlement. B.AI has recently further clarified its product positioning as the "intelligent global settlement layer," meaning it undertakes foundational services including model routing, call metering, and payment settlement between models and Agents.

Currently, B.AI's user base has surpassed 2.28 million. Since its launch on April 9, the platform's single-day token processing volume has grown from 21.1 million to over 1.1 trillion, achieving the leap from tens of millions to trillions in just 145 days. For comparison, leading global multi-model service platform OpenRouter took approximately 2 years and 9 months to reach a single-day trillion-token scale from its launch. Using this public timeline as a reference, B.AI reached the same throughput level roughly 7 times faster than OpenRouter.

From Tens of Millions to Trillions: B.AI Enters the Era of AI Infrastructure at Scale

For model service platforms, token throughput reflects real usage intensity more accurately than raw registered user numbers. This is especially true in code generation, long-text processing, and Agent workflows, where a single task may involve continuous calls to models, tools, and context. Whether a platform can reliably handle high-frequency, high-concurrency requests directly determines its ability to evolve into a true infrastructure layer.

B.AI's recent growth demonstrates exactly this characteristic. From August 17 to August 31 alone, the platform's single-day token throughput grew approximately 242-fold, processing a cumulative 6.86 trillion tokens during that period while adding over 196,000 new API users. Compared to the early days after launch, when usage was dominated by chat and single model calls, an increasing share of traffic now comes from APIs, continuously running developer tasks, and Agent workflows.

The rapid expansion of model supply is also a key growth driver. Recently, B.AI has been continuously opening up domestic models such as DeepSeek, GLM, Qwen, Kimi, MiniMax, Tencent Hunyuan, and Xiaomi MiMo, while offering free calls for certain popular models. Developers can test different models under a unified API system, switching or combining them based on reasoning capability, context length, response speed, and cost — all without needing to maintain separate accounts and API keys for different providers.

This also makes the comparison between B.AI and OpenRouter increasingly meaningful. Both platforms use multi-model access and unified APIs as key entry points, but B.AI is further extending its capabilities into payments, settlement, and Agent infrastructure. Breaking through one trillion tokens in single-day throughput first validates its model-calling and traffic-bearing capacity, while also providing real usage-scale support for its evolution into deeper AI infrastructure.

From Model Gateway to Global Settlement Layer: B.AI Extends Capabilities into the Agent Economy

If unified model calling solves the question of "what AI to use," then as Agents truly enter大规模 applications, the industry must also address the questions of "how to call, how to pay, how to settle, and how to establish identity." As call volumes continue to expand, B.AI's product布局 around model routing, metering, payment settlement, and Agent infrastructure is beginning to show clearer economies of scale.

B.AI has also been further expanding its infrastructure capabilities recently, extending from unified model access into model routing, call metering, and payment settlement, encapsulating this product direction under the concept of the "intelligent global settlement layer." The logic behind this is that in the future, when an Agent completes a task, it may need to sequentially call different models, data services, on-chain tools, and even other Agents — with each call corresponding to resource consumption and value exchange.

Around this direction, B.AI has already built out two layers of capability: unified API and payment settlement. On one hand, developers can call different models through a single account and API; on the other, the platform supports traditional payment methods including bank cards, WeChat Pay, Alipay, and UnionPay, as well as multiple blockchain networks such as TRON, BNB Chain, and Ethereum, along with various crypto asset deposits. For individual developers and teams across different regions, model calling and payment can both be completed within the same platform.

On the Agent side, B.AI continues to advance capabilities including x402 payments, on-chain identity, Agent Wallet, as well as MCP and Skills, enabling Agents in the future not only to call models and tools but also to autonomously execute payments, settlements, and on-chain operations. Compared to simply offering a list of models, this system is closer to a "foundational services layer" built around AI calls.

From 2.28 million users to a single-day processing scale exceeding 1.1 trillion tokens, B.AI has accumulated a substantial volume of real call traffic. As call volumes continue to grow, the metrics by which the market evaluates such platforms are also shifting. Beyond model count and token throughput, API stability, model routing, payment settlement, and Agent service capabilities are becoming key indicators for measuring a platform's infrastructure strength.