Qwen released Qwen-Image-2.1-Turbo, an accelerated checkpoint of its 7B image generation model that creates and edits images in 8 denoising steps. Open weights are available on Hugging Face, with Diffusers support via QwenImage21Pipeline. Qwen claims the reduced step count does not degrade quality, with strong 2K text-to-image output and natural-language editing retained.
Alibaba's Qwen team released Qwen-Image-2.1 as open weights, a unified 7B model for both image generation and editing. It natively generates and edits RGBA transparent images, supports up to 10 reference images, and targets panoramas, infographics, and virtual try-ons. Day-0 support arrived from vLLM and ComfyUI, plus a live Hugging Spaces demo, giving practitioners an immediately usable open alternative to closed-source image models.
Alibaba's Qwen team released Qwen-Image-2.1 as open weights, a unified 7B model for both image generation and editing. It natively generates and edits RGBA transparent images, supports up to 10 reference images, and targets panoramas, infographics, and virtual try-ons. Day-0 support arrived from vLLM and ComfyUI, plus a live Hugging Spaces demo, giving practitioners an immediately usable open alternative to closed-source image models.
The maintainer of TensorSharp, an open-source inference engine, added support for running Qwen-Image 2.1 locally for text-to-image generation and image editing, including LoRA adapters. The update ships GGUF model configs plus accelerated adapters like Pruna 8-step and Viggle Turbo that cut sampling passes. It matters to practitioners running image models entirely on local hardware without vendor runtimes.
The maintainer of TensorSharp, an open-source inference engine, added support for running Qwen-Image 2.1 locally for text-to-image generation and image editing, including LoRA adapters. The update ships GGUF model configs plus accelerated adapters like Pruna 8-step and Viggle Turbo that cut sampling passes. It matters to practitioners running image models entirely on local hardware without vendor runtimes.
The developer of Kokon, a no-settings desktop editor for Qwen-Image 2.1, released version 0.10 with two community-requested features: Prompt+, which expands short ideas into detailed prompts using Qwen3-VL 8B (the text encoder the app already loads, so no extra download), and beta support for loading users' own Qwen-Image 2.1 LoRAs. Kokon is free, AGPL-licensed, and now runs a custom stable-diffusion.cpp build on Windows, macOS, and Linux.
The developer of Kokon, a no-settings desktop editor for Qwen-Image 2.1, released version 0.10 with two community-requested features: Prompt+, which expands short ideas into detailed prompts using Qwen3-VL 8B (the text encoder the app already loads, so no extra download), and beta support for loading users' own Qwen-Image 2.1 LoRAs. Kokon is free, AGPL-licensed, and now runs a custom stable-diffusion.cpp build on Windows, macOS, and Linux.
The developer of TensorSharp, an open-source local AI inference engine, announced support for Qwen Image 2.1 Turbo and LoRA. A demo video shows region-based image editing: users mask an area and modify it with natural-language prompts. The project aims to unify local inference for LLMs, multimodal models, and image generators.
The developer of TensorSharp, an open-source local AI inference engine, announced support for Qwen Image 2.1 Turbo and LoRA. A demo video shows region-based image editing: users mask an area and modify it with natural-language prompts. The project aims to unify local inference for LLMs, multimodal models, and image generators.
Intel announced Day-0 OpenVINO optimization for Qwen-Image-2.1, enabling the open-weight image model to run optimized on Intel hardware. The single checkpoint supports both image generation and editing, so developers can deploy one artifact for both tasks.
Intel announced Day-0 OpenVINO optimization for Qwen-Image-2.1, enabling the open-weight image model to run optimized on Intel hardware. The single checkpoint supports both image generation and editing, so developers can deploy one artifact for both tasks.