Model positioning
Qwen-Plus-Character is a text generation and role-playing model in the Qwen series, designed for anthropomorphic dialogue. It is a dynamically updated version, and the official documentation states that model updates will be announced in advance. (docs.bailian.console.aliyun.com)
Capabilities and uses
The model is optimized for following instructions within a defined persona, faithfully portraying characters, advancing conversations, and listening and showing empathy. Use cases include virtual social interaction, game NPCs, recreation of IP characters, and interaction with smart hardware such as toys and in-vehicle systems. Developers can specify character backgrounds, relationships, scenarios, and language styles in system messages. (docs.bailian.console.aliyun.com)
Limitations
- Both input and output modalities are text. The official model information page indicates support for structured output and context caching, but not Function Calling, prefix continuation, batch inference, or model tuning. Web search support varies by region: it is supported in Beijing but not in Singapore. (help.aliyun.com)
- In the Beijing region, the context length, maximum input, and maximum output are each 32,768 Tokens. In the Singapore region, the context length and maximum input are 131,072 Tokens, the maximum output is 32,768 Tokens, and the default output limit is 4,096 Tokens, adjustable through
max_tokens. Input and output must still comply with the total context length limit. (help.aliyun.com) - The dynamic version's performance may change with updates; the official documentation at the time of use should be consulted for region-specific capabilities and limitations. (docs.bailian.console.aliyun.com)
Pricing
The following are the original API usage prices listed on the official pages, all per million Tokens, as checked on 2026-11-19. Prices for different sites and regions should not be regarded as direct currency conversions of the same plan.
- Alibaba Cloud China site, Chinese yuan (CNY): In the Beijing region, input costs 0.8 yuan and output costs 2 yuan; in the Singapore region, input costs 3.747 yuan and output costs 10.492 yuan. Input cache hits cost 0.16 yuan in Beijing and 0.749 yuan in Singapore; explicit cache creation costs 1 yuan and explicit cache hits cost 0.08 yuan in Beijing. Pricing source: Alibaba Cloud model information and pricing. (help.aliyun.com)
- Alibaba Cloud international site, US dollars (USD): In the Singapore region, input costs 0.5 USD and output costs 1.4 USD; in the Beijing, Virginia, US, Frankfurt, Germany, Tokyo, Japan, and Hong Kong, China regions, input costs 0.115 USD and output costs 0.287 USD. Tokens for session cache hits are billed under the implicit caching rules; see official pricing for the applicable rates. Pricing source: Alibaba Cloud role-playing models and pricing. (alibabacloud.com)
Release date and official announcement
The date in the official listing record is 2025-03-20: Alibaba Cloud Model Studio model listings and updates. (docs.bailian.console.aliyun.com)