Changes

Model Studio will start to update and upgrade the stable versions of Qwen-Plus and Qwen-Max on September 16, 2025.

In both Beijing and Singapore region, qwen-plus will update to qwen-plus-2025-07-28. Compared with the previous version, its Chinese, English, and tool calling capabilities are specially enhanced. It supports 1 million context window for the first time, with tiered pricing by context length.

Impact

After the update:

Beijing region:

• qwen-plus will update to qwen-plus-2025-07-28.
  • The prices of the stable model will be aligned with the snapshot. Cached prices will change accordingly.
  • The max_input_tokens defaults to 128K. Tiered pricing will be used. (Note: snapshots prior to qwen-plus-2025-07-28 do not have context-length tiering.)
  • 30 days later, the system will automatically adjust the default value of max_input_tokens from 128K to 1M.

Singapore region:

• qwen-plus will update to qwen-plus-2025-07-28.
  • The prices of the stable model will be aligned with the snapshot. Cached and batch prices will change accordingly.
  • The max_input_tokens defaults to 256K. Tiered pricing will be used. (Note: snapshots prior to qwen-plus-2025-07-28 have a maximum context window of 128K, and do not have context-length tiering.)
  • 30 days later, the system will automatically adjust the default value of max_input_tokens from 256K to 1M.

Model mapping

If your business is better suited to the current model version, you can switch to the current version before September 16, 2025.

If you have any questions, submit a ticket to contact us.

Beijing region

• qwen-plus: Current: qwen-plus-2025-04-28; After update: qwen-plus-2025-07-28

Singapore region:

• qwen-plus: Current: qwen-plus-2025-04-28; After update: qwen-plus-2025-07-28

Log on to Model Studio. For model details, see Documentation.