
Daily Token Throughput Surpasses 13.3 Trillion! B.AI Builds an "AI Power Grid" with Full-Stack Infrastructure, Solidifying the "Global Settlement Layer" for the Agent Economy
Recently, the new-generation AI infrastructure platform B.AI launched its "Inclusive Large Model Compute" initiative, triggering an unprecedented wave of large model invocations across developers worldwide. Within just a few days, B.AI's full-platform daily token throughput historically broke through the 1.33 trillion mark, with daily API calls exceeding 10.86 million. Over the past 15 days, B.AI's core throughput achieved an explosive 289-fold growth.
The achievement of this milestone data marks the full-scale ignition of B.AI's "inclusive compute" initiative, yet it is merely the first step in B.AI's strategic blueprint. B.AI's ultimate strategic vision is to completely transcend the role of a mere compute distribution channel and comprehensively build a core hub foundation known as the "Intelligent Global Settlement Layer," thereby truly supporting massive cross-node collaboration, scheduling, and value allocation for countless Agents in complex business operations.
As the AI industry fully enters the era of high-frequency Agent collaboration, facing increasingly massive concurrency demands, various underlying large models resemble isolated "power plants" scattered everywhere, while B.AI becomes the "AI Grid" that interconnects the entire era. With forward-looking vision, B.AI precisely positions itself "Above All Models, Below All Agents." By deeply integrating diverse top-tier models and full-stack components, it has forged the most robust and irreplaceable infrastructure foundation to support the large-scale deployment of Agents and the explosion of productivity.
Daily Throughput Breaks 1.33 Trillion! B.AI "Inclusive Compute" Campaign Ignites Platform-Wide Invocation Frenzy
Recently, B.AI's campaign offering free unlimited access to large models rapidly gained traction within the developer community, becoming an exceptionally hot focal topic in the industry. This highly-inclusive action, akin to a "compute carnival," not only brought remarkable activity to the platform but also set a series of striking milestone figures.
Within just a few days, invocation volume on the B.AI platform surged, driving fullplatform daily token throughput to forcefully break 1.33 trillion, with a 15-day cumulative token throughput reaching 8.19 trillion, and cumulative new API users exceeding 220,000.As of September 3, B.AI's total platform users had forcefully surpassed 2.3 million. This traffic wave triggered by "inclusive compute" is continuously refreshing usage records for AI infrastructure with the most authentic market response.
Such a massive scale of invocations directly benefits from B.AI's highly sincere and precisely targeted zero-cost matrix that hits developers' "cost anxiety" pain points. During this campaign, B.AI fully opened a lineup of 6 cutting-edge top-tier models, providing the entire network with a "zero threshold, unlimited" free access experience.

This production-grade matrix precisely covers the full-scenario needs of advanced applications: GLM-5.3-Flash (Ox Alpha) and DeepSeek-V4-Flash-Vision-Exp strongly support multimodal vision and ultra-long text parsing; DeepSeek-V4-Flash paired with Qwen3.8-Flash aligns with ultra-high concurrency response and complex code logic generation; complemented by Tencent Hy3, which excels in Chinese foundational capabilities, and Xiaomi MiMo-V2.5, adapted for all scenarios. From rapid text processing and code reasoning to advanced visual parsing, B.AI uses this top-tier lineup to thoroughly clear away cost barriers for developers pursuing complex AI applications.
It is precisely this "zero-threshold" top-tier model matrix that directly catalyzed the surge in developer adoption of B.AI. Previously, exorbitant token bills directly restricted the deployment of complex Agent architectures. Now that B.AI has completely leveled the compute cost playing field, developers can finally freely deploy Agent architectures capable of long-chain tasks such as "autonomous planning, tool invocation, and self-reflection/error-correction" on a large scale into real-world production and testing environments.
Notably, starting September 3, B.AI adjusted the pricing for the DeepSeek-V4-Flash and DeepSeek-V4-Flash-Vision-Exp models, introducing tiered discounts (50% off during peak hours, down to 25% of official peak prices during off-peak periods), while retaining the free access mechanism for GLM-5.3-Flash (Ox Alpha), Qwen3.8-Flash, Hy3, and MiMo-V2.5. Amidst this commercialization transition, developer invocation enthusiasm remains undiminished, with the full-platform token throughput maintaining a steady and robust upward momentum.
This fully demonstrates that this viral platform-wide campaign was not merely a concentrated release of compute dividends, but also a perfect "barometer" for the evolution of industry infrastructure. It conclusively proves with the real-world scenario of 1.33 trillion tokens per day that AI applications have officially bid farewell to the simple "Q&A era," and B.AI, backed by a solid technological foundation and refined service mechanisms, is comprehensively unlocking deep empowerment for the large-scale deployment of advanced AI Agents.
B.AI Builds an "AI Grid," Consolidating the "Global Settlement Layer" for the Agent Economy
The "inclusive compute" initiative crafted by B.AI is merely the starting point, not the end goal. For B.AI, this "compute carnival" serves even more as an extreme stress test for the platform's underlying capabilities. When massive numbers of developers flood in simultaneously, and long-chain Agent tasks flow across the platform at unprecedented densities, the true challenge comes to light: how to ensure this burst transcends a single event and instead solidifies into a sustainable, evolvable, and billable systemic capability? At this pivotal juncture, B.AI chooses to deepen its strategic focus, dedicating itself to building full-stack infrastructure that powers the agent-driven era, thereby consolidating its hub positioning as the "Intelligent Global Settlement Layer."
In traditional understanding, the significance of infrastructure lies in "stable supply." But in the Agent era, the connotation of infrastructure is undergoing a qualitative shift. A typical agent task often requires repeatedly switching between multiple models: using a vision model to understand input, a reasoning model to plan steps, a code model to generate executable logic, and then a lightweight model to complete formatted output. This "multi-model relay" workflow dictates that no single model vendor can independently support a complete Agent workflow. Developers are forced to confront the fragmented reality of dealing with multiple API protocols, multiple billing systems, and multiple rate-limiting rules.
The settlement layer constructed by B.AI precisely bridges this fracture. With forward-looking vision, B.AI precisely positions itself "Above All Models, Below All Agents," abstracting models from different vendors, with varying capabilities and cost structures, into a schedulable resource.
Relying on a tiered API system characterized by "official stability guarantees and self-selected lowest-priced options,"developers can secure deterministic assurance via direct official connections while flexibly selecting premium supplies across a rich model array with varying discount tiers, executing tasks of equal quality at more competitive costs. Coupled with the "smart routing" feature on the Chat end, B.AI has created a full-stack "AI Grid,"ensuring that every task request from an Agent lands precisely on the most suitable model at a reasonable cost with stable performance.

Furthermore, the settlement capability of this grid equally accommodates the dual logic of Web2 and Web3. On the Web2 side, developers can continue using familiar traditional payment methods to recharge and settle bills with minimal friction, seamlessly integrating into existing business workflows. On the Web3 side, B.AI leverages on-chain payment channels to offer decentralized, verifiable, and low-friction payment options for developers worldwide.
The parallel implementation of these two payment systems enables developers and agent applications from diverse backgrounds and infrastructure preferences to find their most suitable settlement path on this grid, truly realizing "one-time integration, global circulation."
Targeting Core Productivity, B.AI's Full-Stack Infrastructure Reshapes the Agent Collaboration Network
After establishing a solid compute routing and global settlement foundation, B.AI further bridged the gap from infrastructure deployment to productivity reshaping, comprehensively empowering agents to achieve high-efficiency collaboration across diverse business scenarios. This completes the most critical execution-layer puzzle for the comprehensive breakout of the agent economy.
The core of this puzzle is its access support for Codex.As the core tool for code execution and the agent action layer, Codex holds absolute technical advantages in understanding project architecture, proactively planning, and executing complex development tasks. Currently, the B.AI API is fully compatible with the Responses API. Developers can use their B.AI API Key to invoke GPT-series models supported by B.AI, as well as the highly favored DeepSeek-series models, directly within Codex.

Through this seamless integration, technical teams can embed these two top-tier code computing powers directly into daily coding workflows, meeting multi-dimensional needs such as code generation, logic comprehension, bug debugging, system refactoring, and development collaboration in one stop. This provides developers with a complete closed-loop experience from model selection to code deployment.
Achieving seamless docking between compute and application layers is merely the first step in unleashing large model productivity. Beyond this, realizing the agent economy requires breaking down the collaborative barriers among diverse Agents under task chains. To enable massive Agents to operate stably in real-world production environments, B.AI has constructed a full-stack infrastructure comprising five core components, equipping agents with an extremely precise operational system:
- x402 Payment Protocol: Unlocks automatic micro-payments between Agents. Built on the HTTP 402 standard, this protocol innovatively introduces a "pay-before-response" mechanism, enabling seamless high-frequency on-chain automatic micro-payments and settlements in the background when Agents perform cross-API calls or compute collaboration.
- 8004 Identity Authentication Protocol: Forges exclusive "digital ID cards" for Agents. The 8004 protocol generates objective, verifiable identity credentials for each Agent on-chain, archiving their call history and credit scores, thus completing the crucial credit closed loop for cross-node collaboration in the intelligent economy.
- Skills "Skill Bricks": Equips agents with out-of-the-box toolchains. Working in deep coordination with MCP Servers, standardized Skills matrices grant Agents highly modular tool invocation capabilities, thoroughly bridging the last mile of intelligence from "thought and decision-making" to "action execution."
- BAIclaw and BAIcode: Native super assistants of the platform. As super-application forms targeting end-users and developers, BAIclaw and BAIcode represent the ultimate embodiment of B.AI's underlying architecture and scheduling capabilities. Specifically, BAIclaw focuses on all-encompassing smart terminals and multi-agent collaboration, deeply integrating top-tier large model matrices. Users simply input natural language commands to automatically drive Agents to complete complex on-chain economic activities and provide all-around-the-clock management. As an advanced AI programming tool, BAIcode combines the top-tier intelligence of large models to achieve a full-process closed loop from requirement comprehension and architecture planning to code generation and testing.
From the surge in token throughput sparked by the free-access frenzy, to the architectural establishment of the "Intelligent Global Settlement Layer"; from the seamless Codex integration, to the full-stack infrastructure built upon x402, 8004, Skills, and native assistants, what B.AI presents is not merely a hardcore technical solution, but a clear futuristic blueprint. In this AI-driven industrial reshaping, B.AI consistently and firmly positions itself "Above All Models, Below All Agents," transforming fragmented and dispersed compute supply into a highly available, fluid, and billable system-level network capability.
As tens of millions of Agents seamlessly flow, freely collaborate, and settle in milliseconds on this infrastructure, AI is rapidly bidding farewell to its tool-like attribute of one-way responses, truly evolving into an economic entity driving leaps in social productivity. What B.AI consolidates is not only the present state of inclusive compute, but the ultimate foundation leading to the massive breakout of the agent economy. The future is here, and this is only the beginning.














