On Monday, Alibaba (BABA) unveiled a new AI model called Qwen3.5, aimed at executing complex tasks independently, with improvements in performance and cost.
The Chinese ecommerce giant said that in various task evaluations, the 3.5 series consistently demonstrates performance on par with state-of-the-art leading models. “Compared to the 3 series, these models show a leap forward in both pure-text and multimodal capabilities. This version is functionally equivalent to snapshotqwen3.5-plus-2026-02-15,” said the company.
The company noted in a blog post that it is open-sourcing Qwen3.5-397B-A17B, also named Qwen3.5-Plus.
Qwen3.5 can process text, image and video, and generate text, according to the company.
Qwen3.5-397B-A17B shows strong performance across language understanding and reasoning, code generation, agentic workflows, image and video comprehension, and GUI interaction, rivaling leading frontier models in both versatility and performance, as per the company.
Alibaba said that the model delivered up to a 60% reduction in deployment expenses while matching the performance of the much larger Qwen3-Max model, which exceeds one trillion parameters.
“Built for the agentic AI era, Qwen3.5 is designed to help developers and enterprises move faster and do more with the same compute, setting a new benchmark for capability per unit of inference cost,” said the company, Reuters reported.
Alibaba’s AI models face tough competition in the Chinese market from models built by companies such as ByteDance (BDNCE), Baidu (BIDU), and DeepSeek (DEEPSEEK).
On Saturday, ByteDance released its AI model Doubao 2.0, which, according to the company, is positioned for the “agent era.” The company’s recently released AI video-generation Seedance 2.0 is drawing praise overseas on platforms such as X, including from Elon Musk, the report added.
DeepSeek is also expected to release its new AI model in the coming days.
Alibaba did not mention DeepSeek in its announcement for Qwen3.5, and the many benchmarks it published only show the new model outperforming a previous iteration and U.S. models GPT-5.2 from OpenAI (OPENAI), Claude Opus 4.5 from Anthropic (ANTHRO), and Gemini 3 Pro from Google, a unit of Alphabet (GOOG) (GOOGL).