The latest AI news and updates of Qwen — Alibaba's family of open-source and commercial models spanning text, vision, and code. Covering Qwen's latest launches, product updates, and company news from the past 90 days.
QwenJun 1Alibaba released Qwen3.7-Plus, a multimodal foundation model that integrates vision and language reasoning into a single agent loop. The update enables AI agents to perceive graphical interfaces and execute command-line operations to automate complex software development and productivity workflows.
QwenMay 27Qwen achieved a record 580 tokens per second running its Qwen3.5-397B-A17B model on NVIDIA Blackwell GPUs using the TokenSpeed inference engine. The optimization targets agentic workloads, where multi-turn reasoning and tool-calling typically suffer from high latency. By combining a hybrid attention architecture with deep kernel fusion, the system maintains high throughput even as context scales to one million tokens.
QwenMay 25Qwen introduced implicit and explicit context caching for its flagship Qwen3.7-Max model to reduce API latency and expenses. By allowing developers to pin massive system prompts and tool definitions, the update cuts the cost of repeated inputs by 90 percent.
QwenMay 21Alibaba released Qwen3.7-Max, a flagship model optimized for autonomous agents capable of executing multi-step tasks over dozens of hours. The model features native support for the Model Context Protocol and demonstrated a tenfold performance increase in self-directed kernel optimization.
QwenMay 1Alibaba's Qwen team partnered with Fireworks AI to provide production-ready access to its closed-weights Qwen 3.6 Plus model. This move gives global developers a low-latency, cost-effective way to run Alibaba's flagship intelligence without using Chinese cloud infrastructure.
QwenApr 30Alibaba's Qwen team launched Qwen-Scope, an open-source suite of sparse autoencoders that map the internal neural patterns of the Qwen3 and Qwen3.5 model families. By converting dense hidden layers into interpretable features, the toolkit allows developers to steer model behavior, synthesize targeted training data, and debug generation errors at the root.
QwenApr 30Alibaba's Qwen team launched FlashQLA, a library of high-performance linear attention kernels that deliver up to a 3x speedup for on-device models. By optimizing the Gated DeltaNet architecture for personal hardware, the release makes long-context agentic workflows more viable on consumer electronics.
QwenApr 25Alibaba released Qwen-Image-2.0-Pro, a unified model that merges image generation and editing into a single 7B-parameter architecture. The update introduces 1K-token instruction support and native 2K resolution, shifting AI imagery from artistic illustration toward functional, text-heavy professional design.
QwenApr 24Alibaba open-sourced Qwen3.6-27B, a dense multimodal model that surpasses the previous 397B-parameter flagship on every major agentic coding benchmark. The release proves that optimized mid-sized dense models can deliver frontier-level performance while remaining significantly easier to deploy than massive mixture-of-experts architectures.
QwenApr 20Qwen released an early preview of Qwen3.6-Max-Preview, its next flagship model designed for complex autonomous workflows. The update delivers a nearly 10-point jump in agentic coding benchmarks and introduces a feature to preserve reasoning traces across multi-turn interactions.
QwenApr 16Alibaba open-sourced Qwen3.6-35B-A3B, a sparse Mixture-of-Experts model with 35 billion total parameters that only activates 3 billion during inference. Despite its small active size, it matches frontier models in agentic coding and multimodal reasoning while operating under a permissive Apache 2.0 license.
QwenApr 10Qwen Code v0.14.0 introduces remote execution via chat apps and automated task scheduling for terminal-based AI agents. These updates shift the tool from a reactive assistant to an autonomous background operator capable of managing long-running engineering workflows.
QwenApr 9Alibaba Qwen has moved its Qwen3.6-Plus model out of trial and into full production availability on platforms like OpenRouter and Fireworks AI. The model delivers frontier-level reasoning and a 78.8 SWE-bench score, offering a high-performance alternative for repository-level coding at a lower price point.
QwenApr 2Alibaba released Qwen3.6-Plus, a frontier model designed for autonomous agentic workflows across text, image, and video. The update marks a shift toward native multimodality, enabling agents to reason through visual data and execute complex multi-step coding tasks.
QwenMar 31Alibaba Cloud released Qwen3.5-Omni, a native omni-modal foundation model that processes text, images, audio, and video within a single end-to-end architecture. This update introduces real-time interactive capabilities and a new feature called Audio-Visual Vibe Coding.
QwenMar 31Alibaba has launched an early preview of Qwen 3.6 Plus on OpenRouter, featuring an advanced hybrid architecture. This update specifically targets the reliability of function calling and tool selection, making it a specialized choice for developers building autonomous coding agents.
Frequently asked questions