Notice on decommissioning selected legacy Qwen3.5 models on January 19, 2027, with the full list and recommended replacements.
Affected time
2027-01-19 10:00:00 (UTC+08) [Subject to the actual change time]
Change details
To continuously enhance model performance and service experience while optimizing underlying resource management, QwenCloud will decommission the legacy models listed in the table below on January 19, 2027 (UTC+8).
The detailed retirement list and recommended replacement models are as follows:
| Model name | Input cost ($/million tokens) | Output cost ($/million tokens) | Recommended replacement model | Input cost ($/million tokens) | Output cost ($/million tokens) |
|---|---|---|---|---|---|
| qwen3.5-plus | 0.4 | 2.4 | qwen3.8-flash | 0.15 | 0.47 |
| qwen3.5-plus-2026-02-15 | 0.4 | 2.4 | qwen3.8-flash | 0.15 | 0.47 |
| qwen3.5-plus-2026-04-20 | 0.4 | 2.4 | qwen3.8-flash | 0.15 | 0.47 |
| qwen3.5-397b-a17b | 0.6 | 3.6 | qwen3.8-flash | 0.15 | 0.47 |
Please be advised that QwenCloud will continue to undergo rapid iterations and release higher-performing models between the publication of this notice and the actual retirement dates. Consequently, the replacement recommendations provided in this notice will not be frequently updated with the release of new models. We highly recommend regularly visiting the Models page during your migration and daily operations to ensure you can flexibly select the latest and most suitable models for your specific business needs.
Action required
Key milestones
To ensure a smooth transition for your business, please note the following two key dates:
- November 18, 2026 (End of Marketing - EOM): The purchase portal for the aforementioned models will be closed. You will no longer be able to activate new subscriptions or purchase related services and resource packs for these models.
- January 19, 2027 (Official Decommissioning): The models will completely cease inference services.