Qwen: Qwen3.5 Plus 2026-02-15

The Qwen3.5 native vision-language series Plus models are built on a sophisticated hybrid architecture that integrates linear attention mechanisms with sparse mixture-of-experts models. This design choice significantly enhances inference efficiency, allowing for faster and more cost-effective processing of complex tasks. In a variety of comprehensive task evaluations, the 3.5 series consistently demonstrates performance on par with state-of-the-art leading models, setting a new benchmark for AI capabilities. Compared to the previous 3 series, these models show a remarkable leap forward in both pure-text and multimodal capabilities. Key features include json_mode for structured output, streaming for real-time interactions, functions for tool use, and long_context (1000K tokens) for extensive understanding. It also supports video_in and vision for advanced multimodal processing. Max output is 65K tokens. Pricing is competitive at $0.40/2.40 per 1M tokens (input/output), available on our PRO Access Tier. Multi AI provides seamless access to Qwen3.5 Plus, allowing developers and businesses to leverage its advanced features for a wide range of applications, from complex data analysis to innovative content generation. Experience the next generation of AI with Qwen3.5 Plus.