Affected time
2026-05-13 00:00:00 (UTC+08) (Depending on the actual time)
Changes
To continuously enhance model performance and service experience, while optimizing underlying resource management, QwenCloud will decommission the snapshot models for the Beijing and Singapore regions listed below on May 13, 2026.
The specific decommissioning list is as follows:
- qwen-max series snapshots: qwen-max-latest, qwen-max-2025-01-25
- qwen-turbo series models: qwen-turbo-latest, qwen-turbo-2025-04-28, qwen-turbo-2024-11-01
- qwen-vl-max series snapshots: qwen-vl-max-latest, qwen-vl-max-2025-08-13, qwen-vl-max-2025-04-08
- qwen-vl-plus series snapshots: qwen-vl-plus-latest, qwen-vl-plus-2025-08-15, qwen-vl-plus-2025-07-10, qwen-vl-plus-2025-05-07, qwen-vl-plus-2025-01-25
- qvq-max series snapshots: qvq-max-latest, qvq-max-2025-03-25
- Open-source series snapshots: qwen2.5-vl-32b-instruct, qwen2.5-vl-72b-instruct, qwen2.5-vl-7b-instruct, qwen2.5-vl-3b-instruct, qwen2.5-7b-instruct-1m, qwen2.5-14b-instruct-1m, qwen2.5-72b-instruct, qwen2.5-32b-instruct, qwen2.5-14b-instruct, qwen2.5-7b-instruct, qwen3-0.6b, qwen3-1.7b, qwen3-4b
Recommended alternatives
We recommend replacing the soon-to-be-decommissioned models with the latest Qwen3.6 series models.
The Qwen3.6 series demonstrates exceptional performance comparable to current state-of-the-art frontier models, delivering significant overall improvements over the Qwen3.5 series and earlier historical versions. Specifically, the Qwen3.6-Plus model features notable enhancements in:
- Code development capabilities: Agentic Coding, front-end programming, Vibe Coding, and more;
- Visual understanding capabilities: Multi-modal universal recognition, high-precision OCR, fine-grained object localization, and more.
Given that QwenCloud will continue to rapidly iterate and release new, higher-performance models between the announcement date and the actual decommissioning date, the replacement recommendations outlined in this notice will not be frequently updated to reflect newly launched models. We kindly encourage you to stay tuned to the Models page during your business migration and daily usage, so you can flexibly adopt the latest model versions best suited to your specific business scenarios.
Impact of the change
After decommissioning, model inference will no longer be supported for these models. Applications and services that continue to call the decommissioned models will fail to return results.
If you are currently using any of the above models, please ensure that you complete the following before May 13, 2026:
- API endpoint migration: Update your integration to point to the recommended replacement models;
- Parameter tuning: Adjust inference parameters as needed for the new models;
- Business validation: Thoroughly test and verify your workflows to ensure expected performance.