Model Positioning
The Plus model in the Qwen3 series is designed for text generation and reasoning. It combines thinking and non-thinking modes, which can be switched during a conversation. This version is the July 14, 2025 snapshot; compared with the previous version, the official announcement states that its Chinese and English capabilities in non-thinking mode have improved and its tool-calling capabilities have received targeted enhancements. (docs.modelstudio.console.alibabacloud.com)
Capabilities and Uses
- Supports text input and text output, and can be used for Chinese and English conversations, text generation, and tasks requiring reasoning.
- Alibaba Cloud Model Studio's Beijing region supports function calling, structured output, web search, and prefix completion, making it suitable for building applications that use tool calling or formatted output.
- Capabilities vary by region: Model Studio's Singapore international deployment supports prefix completion, but function calling, structured output, and web search are marked as unsupported for this version. (help.aliyun.com)
Limitations
- The context window is 131,072 tokens; the maximum input length is 98,304 tokens, and the maximum output length is 16,384 tokens.
- Both the Beijing and Singapore deployments of Alibaba Cloud Model Studio mark context caching, batch inference, and model tuning as unsupported for this version. Features such as web search should be verified for the actual deployment region.
- The official rate limits listed for both the Beijing and Singapore regions are 60 requests per minute and 100,000 tokens per minute; actual usage is subject to the platform's current configuration. (help.aliyun.com)
Release Date
An Alibaba Cloud Model Studio announcement lists the model in the release entry dated July 16, 2025; July 14, 2025 is the model snapshot date. The international deployment announcement lists it in the July 23, 2025 entry. (docs.modelstudio.console.alibabacloud.com)
Pricing
The following are the officially published unit prices for this specific snapshot. Input and output are billed separately, all prices are per million tokens, and there are no input-length tiers; output billing in thinking mode includes the chain of thought and the answer. Date checked: November 18, 2026. Prices and availability may change; actual billing is subject to the relevant platform. (help.aliyun.com)
| Platform and region | Currency | Non-thinking input | Non-thinking output | Thinking input | Thinking output |
|---|---|---|---|---|---|
| Alibaba Cloud Model Studio China site, China North 2 (Beijing) | CNY | 0.8 yuan | 2 yuan | 0.8 yuan | 8 yuan |
| Alibaba Cloud Model Studio China site, Singapore international deployment | CNY | 2.936 yuan | 8.807 yuan | 2.936 yuan | 29.357 yuan |
| QwenCloud international site, Singapore international deployment | USD | 0.4 dollars | 1.2 dollars | 0.4 dollars | 4 dollars |
The above are official prices for different sites and deployment conditions, not RMB prices converted to US dollars at an exchange rate. The Model Studio model information page states that the listed prices are standard invocation prices and exclude limited-time discounts. (help.aliyun.com)