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 analysis from the past 90 days.
WanOct 20Alibaba's Wan 3.0 Prime and Wan 3.0 models secured the #1 and #3 spots on OpenRouter's New & Trending video leaderboard. Wan 3.0 Prime reached the top rank with 3,000 requests and an 85% increase, while Wan 3.0 holds the third position with 29,000 requests and a 52% increase, based on usage data through October 19, 2026.
QwenOct 12Qwen introduced QwenGyre, an elastic reinforcement learning framework for training extreme-long-horizon agents. It features an Elastic Scheduler that reallocates GPUs without interrupting live tasks and a Trajectory Processor that deduplicates branching execution paths. Scaled to Qwen3.8-2.4T-A95B, it lifts NL2RepoBench from 52.5% to 58.5% in 48 steps and cuts end-to-end training time by 44% versus Async.
QwenOct 9Alibaba released Qwen-Image-2.1-Turbo, an accelerated open-weights checkpoint of its 7B visual generation model. The model produces 2K images and supports natural-language editing in just 8 denoising steps. It integrates with the Diffusers library via the QwenImage21Pipeline, and both Turbo and Pro versions are now available through hosted APIs on Alibaba Cloud Model Studio.
QwenSep 23Qwen launched Qwen Intelligence, a mobile agent platform featuring three specialized agents: a Mobile Planner Agent for task orchestration, a Mobile-Use Agent for API and GUI execution, and a Mobile Creative Agent for image generation. The release includes an open benchmark suite covering planning, cross-app execution, and safety. Subscriptions start at ¥78 per month for the standard tier.
QwenSep 23Qwen launched Qwen-Audio-3.1, a five-model audio stack featuring upgraded ASR, TTS, and Realtime capabilities alongside two new models: ASR-Next for audio understanding and TTS-Next for audio creation. The release includes significant price reductions across the lineup, with TTS down about 70%, Realtime down roughly 85%, and ASR up to 95% off.
QwenSep 20Alibaba released Qwen-Image-2.1, an open-weights model that unifies text-to-image generation and editing in a single 7B checkpoint. The model adds native RGBA transparency for generating and editing transparent layers, supports up to 10 reference images for complex compositions, and enables precise local editing through masks and annotations. It also improves portrait fidelity and typography aesthetics.
WanSep 20Alibaba added a Peel-Off Sticker skill to its wan.video platform, powered by the Wan 3.0 model. This tool isolates a subject from a provided photo, enabling its insertion into new scenes with generated motion. The skill is now live on the Wan creative platform.
QwenSep 19Qwen released Qwen3.8-LiveTranslate, a real-time simultaneous interpretation model using an Interleave architecture to reduce average lagging from 2.8s to 2.3s across 60 languages. The update adds real-time speaker diarization with voice cloning, synchronized bilingual display, and long-context disambiguation. The model is available via the DashScope API on QwenCloud.
WanSep 18Alibaba's Wan 3.0 model secured second place in the Overall Video category on the OpenArt Arena leaderboard with an overall score of 1,047. The model supports 30-second video generation, native audio, and reference-based inputs. The company shared the ranking alongside its latest performance results on the platform.
QwenSep 18Qwen launched Qwen3.8-Omni-Flash, an omni-modal model for agentic audio-video tasks. It features a 1M-token context window, improves agent performance by 19.5 points on key benchmarks, and reduces video input costs by about 89% compared to Qwen3.5-Omni-Plus. The release includes open-sourced Qwen-MM-Plugins and the Qwen-Live Harness for real-time interaction and tool orchestration.
QwenSep 3Qwen launched E-Commerce Bench, a benchmark evaluating AI agents on 365-day autonomous business operations using real market data. Agents manage inventory, pricing, and supplier negotiations across 6,886 products. The evaluation scores performance across seven axes, revealing that most models struggle to improve strategies over the year, with no single model dominating across all dimensions.
QwenSep 2Alibaba released Qwen3.8-Max-0902, an upgraded model featuring 2.4T parameters and a 1M-token context window. Post-trained on coding and collaborative work, the model debuted at #1 on the Code Arena: WebDev leaderboard with a score of 1,691. It is available via API on QwenCloud, priced at $2 per million input tokens and $6 per million output tokens.
QwenAug 27Alibaba’s experimental Qwen3.8-Flash-Next model now features day-0 support for NVIDIA NeMo AutoModel and NeMo RL, enabling domain-specific fine-tuning. The integration includes recipes for running the model with SGLang, vLLM, and LightSeek TokenSpeed. This release provides an early preview of the upcoming Qwen4 architecture, allowing immediate testing and optimization on NVIDIA hardware.
QwenAug 26Qwen released the Qwen3.8-Flash API on QwenCloud, priced at $0.16 per 1M input tokens and $0.47 per 1M output tokens. This multimodal mixture-of-experts model features a 125B parameter backbone with 6B active parameters and a 1M-token context window. It supports function calling, structured outputs, and context caching for high-concurrency agentic workflows.
QwenAug 26Alibaba released open weights for Qwen3.8-Flash-Next, a multimodal MoE model previewing the Qwen4 architecture. It features a 125B parameter backbone with 6B active parameters, utilizing GDN and QSA hybrid attention for efficiency. The model supports 262K native context, extensible to 1M tokens, and will arrive on QwenCloud at $0.16 per 1M input tokens and $0.47 per 1M output tokens.
WanAug 24Alibaba released Wan 3.0, its latest video generation model, now generally available on Alibaba Cloud Model Studio and Qwen Cloud. The model supports 30-second clips, omni-reference inputs, and document parsing. Standard API pricing starts at $0.05 per second for 480p, with a 30% launch discount available through September 23.
QwenAug 22Alibaba Qwen’s Qwen3.8-27B is now the #1 open-weight model on Harvey’s Legal Agent benchmark, achieving 11.3 points. It ties with Fable 5 and outperforms Kimi K3, Qwen 3.8 Max, and DeepSeek V4. This 27-billion-parameter model delivers professional-grade agentic performance while remaining small enough to run on local hardware.
QwenAug 22Qwen amplified a community comparison showing its 27B open-weight model outperforming Google’s Gemini 3.7 Flash on a one-shot water-surface simulation. The model, running locally on an RTX 3090 at Q5 quantization, produced visually consistent results from a single prompt. Qwen now invites users to run the model on their own laptops.
QwenAug 15Alibaba released Qwen3.8-Max, a 2.4-trillion-parameter model with 95 billion active parameters and a 1-million-token context window. The model is now available for dedicated inference through Day 0 launch partners Together AI, Fireworks AI, and Nebius Token Factory. It supports complex autonomous coding, research, and long-horizon agentic tasks.
WanAug 14Wan has launched Agent Skills, enabling the creation and sharing of custom AI agent workflows for video generation. The platform supports reusable tools like drama or music video generators. A reward campaign offers up to 450 Inspiration Points for published skills through August 31.
QwenAug 10Alibaba released Qwen-MM-Plugins, an Apache-2.0 repository of skills and MCP servers that turn existing agent harnesses into multimodal-native systems. The plugins enable agents to read images, videos, and documents, edit video, and perform 3D or CAD modeling. A guided installer supports integration with harnesses like Claude Code and Codex, allowing agents to handle complex visual and spatial tasks.
QwenAug 6Alibaba's Qwen3.8-Max model now ranks fifth on the Artificial Analysis Intelligence Index and first on the Agentic Index. These independent evaluations measure the 2.4-trillion-parameter model's performance across nine diverse benchmarks, including coding and long-horizon tasks. The rankings confirm the model's top-tier agentic capabilities among 594 evaluated systems.
QwenAug 5Alibaba launched Qwen-Image-3.0-Pro on Qwen Cloud, offering high-fidelity image generation with support for 4.5k-token prompts and text rendering as small as 10px. The model supports 12 languages and 100+ art styles, with pricing starting at $0.04 per image for 1K resolution. It currently ranks second among mainstream models on the Text-to-Image Arena leaderboard.
QwenAug 3Alibaba reached the second position on the Text Arena leaderboard. The proprietary qwen3.8-max model achieved a preliminary score of 1496±10, placing the company behind only the top-ranked lab. This ranking reflects the model's performance across math, coding, and creative writing tasks.
QwenJul 31Alibaba released Qwen-Audio-3.0-ASR-Flash, an automatic speech recognition model featuring improved context consistency, custom hotword support, and structured transcript generation. Internal tests show the model achieves 95.36% recall on medical terminology and 93.24% on industrial terms. The model is available in streaming, file-based, and offline modes through Alibaba Cloud’s Model Studio.
Frequently asked questions