Tencent Cloud Notices > Notice Details
[LLM Service TokenHub] Notice on the Release Progress of Tencent Cloud DeepSeek-V4-Flash GA and Its Vendor Direct Version
2026-08-08 21:51:06
Thank you for your continued attention to and support of Tencent Cloud's large language model services. In light of the actual iteration progress of the model, we have updated the previously released notice regarding DeepSeek-V4-Flash GA. Details are as follows:
1. DeepSeek-V4-Flash GA [Vendor Direct] (Model ID: deepseek-v4-flash-202605) has been released as scheduled
DeepSeek-V4-Flash GA [Vendor Direct] (Model ID: deepseek-v4-flash-202605) was released in sync with the original manufacturer on July 31, 2026, 18:00:00 (Beijing Time), and is now available for normal invocation.
2. The upgrade schedule for DeepSeek-V4-Flash (Model ID: deepseek-v4-flash) has been optimized; existing services remain unchanged
To ensure that the new version is fully validated in terms of both capability and stability, and to help ensure the smooth operation of your business as much as possible, we have optimized the schedule for this upgrade: DeepSeek-V4-Flash (Model ID: deepseek-v4-flash) continues to be served by DeepSeek's DeepSeek-V4-Flash Preview version, and the automatic seamless upgrade mentioned in the previous notice will be postponed, subject to the launch of the new version. Your existing API calls will not be affected — the model's service capabilities, calling method, and billing standards all remain consistent with before, and no action is required on your part.
Meanwhile, given the considerable differences in model capability and performance between the new and previous versions, and in order to avoid any unintended impact on your existing workloads, the GA model will be made available under a new Model ID, allowing you to switch at your own discretion as needed.
3. Launch plan for DeepSeek-V4-Flash-0731 GA (Model ID: deepseek-v4-flash-0731)
We expect to launch the all-new DeepSeek-V4-Flash-0731 GA model (Model ID: deepseek-v4-flash-0731) in the near future, featuring significantly enhanced Agent capabilities and benchmark results far surpassing V4-Pro-Preview.
Thank you for your understanding and support. Tencent Cloud will continue to iterate and enhance our large language model product portfolio to provide you with more powerful model capabilities and services. If you have any questions during use, please submit a ticket for feedback. We will verify and resolve your issue as soon as possible.
If you encounter any issues during use, please feel free to submit a ticket for feedback. We will assist you as soon as possible!
