CloudWalk Technology's Model Hub Adds GLM-5.3, Bringing Its Model Lineup to 72
CloudWalk Technology's Yunqi Model Hub has completed a new round of model capability upgrades, with the latest models including GLM-5.3 and GLM-5.3-Flash now officially available, while Kimi K3 is simultaneously offered at 25% off for API calls.
The platform currently integrates 72 mainstream models, continuously expanding the supply of model services for enterprises and developers.
The newly launched GLM-5.3 series focuses on application scenarios such as coding, long-horizon tasks, and agents.
Among them, GLM-5.3-Flash is offered on the Yunqi platform at 40% off, with an input price as low as RMB 0.48 per million Tokens and an output price of RMB 1.68 per million Tokens; Kimi K3 is currently available for API calls at 25% off, with input and output prices of RMB 15 and RMB 75 per million Tokens, respectively.
Based on recent actual usage on the platform, Kimi K3 has become one of the models with the highest Token call volume on Yunqi, and the call volume of the GLM-5.3 series has also grown rapidly since launch.
Meanwhile, the DeepSeek V4 series, Qwen3 series, MiniMax, and other models are all already available on Yunqi, with some models also enjoying varying levels of price discounts.
As large models continue to iterate, enterprises increasingly need to combine different models according to performance, cost, and task type in their actual business operations.
Yunqi Model Hub brings models from different vendors into a unified service system through a unified Base URL, unified protocol, and unified key, and provides capabilities such as call statistics, fine-grained billing, key and quota management, operation monitoring, and multi-model scheduling, reducing the complexity for enterprises in using and managing multiple models.
The value of Yunqi goes beyond model access.
In the current three-layer product system formed by CloudWalk Technology, Yunqi Model Hub sits at the AI infrastructure layer, together with large models, computing power, training and inference platforms, trusted data spaces, and controlled training grounds, providing continuous AI capabilities for upper-layer businesses; the middle layer is handled by Xingyun Agentic OS and Mofang Agent Lab, responsible for the construction, organization, and collaboration of agents; the top layer forms AI teams oriented to real businesses such as content, marketing, R&D, and corporate functions.
In this system, Yunqi serves as the unified model supply and Token service entry point.
Different AI positions can call suitable models according to the task, and Xingyun further organizes multiple Agents to collaborate in execution, ultimately transforming underlying model capabilities into productivity that can enter enterprise processes.
As new models continue to emerge, Yunqi Model Hub will also continue to expand its model service matrix and further strengthen multi-model scheduling and enterprise-level management capabilities, enabling frontier models to enter AI agents and real business scenarios more quickly.
At present, the GLM-5.3 series and Kimi K3 discount services are both available on the Yunqi platform.
You May Be Interested In
-
2026-10-14CloudWalk Technology has introduced an AI Marketing Team for cross-border e-commerce. Combining AI infrastructure, an agentic operating system and professional growth operations expertise, the business explores a new approach to product marketing and business growth through collaboration among multiple AI agents.
-
2025-02-17Currently, CloudWalk has accumulated over 300 industry knowledge bases, combined with the CongRong Large Model Training and Inference Appliance, providing enterprises with a large number of end-to-end solutions, which is expected to further promote the intelligent upgrade of scenarios such as government affairs and finance.
-
2023-06-09CloudWalk Technology, together with Shenzhen Press Group, showcased a new technological feast for visitors.
