"Alibaba has unveiled Qwen3.5, a new multimodal AI model that the company says is intended to serve as a foundation for digital agents capable of advanced reasoning and tool use across applications. The release reflects the ongoing shift from standalone chatbot deployments toward AI systems that can execute multi-step workflows and operate with minimal human prompting. In a blog post, Alibaba highlighted gains across selected benchmarks, claiming Qwen3.5 outperformed earlier versions and competing frontier systems such as GPT-5.2, Claude 4.5 Opus, and Gemini 3 Pro."
"The company is releasing the open-weight Qwen3.5-397B-A17B model for developers, while a hosted version, Qwen3.5-Plus, will be available through Alibaba Cloud's Model Studio platform. The hosted version includes built-in tool capabilities and an expanded context window of up to one million tokens, aimed at enterprise developers building more complex, multi-step applications. Alibaba also emphasized expanded multilingual support, increasing coverage from 119 to 201 languages and dialects, a move that could appeal to global enterprises operating across diverse markets."
Alibaba unveiled Qwen3.5, a multimodal foundation model designed to enable digital agents with advanced reasoning and tool use across applications. The model demonstrates gains on selected benchmarks, with Alibaba claiming it outperformed earlier versions and competing frontier systems such as GPT-5.2, Claude 4.5 Opus, and Gemini 3 Pro. An open-weight Qwen3.5-397B-A17B model will be released for developers, while a hosted Qwen3.5-Plus on Alibaba Cloud Model Studio will provide built-in tool capabilities and up to one million token context. Multilingual support expanded from 119 to 201 languages and dialects. The release appears amid intensifying competition in China's AI market.
Read at Computerworld
Unable to calculate read time
Collection
[
|
...
]