Alibaba’s Qwen team released a multimodal tool layer within the Qwen-Agent framework, an open-source orchestration layer hosted on GitHub under the QwenLM/Qwen-Agent repository, enabling AI agents to process images, videos, and documents alongside text. The framework integrates tool calling, memory, planning, and native multimodal processing, and supports Retrieval-Augmented Generation (RAG) and code interpretation. The Qwen3.5 series, released in mid-February 2026, was designed for “Native Multimodal Agents,” with vision input support and context windows of up to 1 million tokens. The current flagship, Qwen3.8-Max, released on August 2, 2026, has 2.4 trillion total parameters with 95 billion active, and is optimized for “long-horizon coding and professional workflows.” Alibaba has promised to release open weights for Qwen3.8-Max, following the Qwen3.5 series, which made a 397-billion-parameter variant with 17 billion active freely available.