[enhancement]: Support for Qwen Multimodal Models (Qwen-VL and Image Edit)

[gafda]

Is there an existing issue for this?

• I have searched the existing issues

Contact Details

No response

What should this feature add?

This request is to add native support for integrating and running the Qwen series of multimodal models within InvokeAI, specifically targeting the highly capable models from Alibaba Cloud.

I'm requesting support for:

1. Qwen-VL (Vision-Language): Integration would enable features like advanced image-to-text generation, detailed captioning, visual question-answering, and general image analysis. This could be used to enhance prompt generation.
2. Qwen Image Edit / Agent: Support for Qwen's instruction-driven image editing capabilities. This would allow users to perform complex, natural-language-guided image manipulations and stylistic modifications without relying solely on traditional in-painting masks.

Alternatives

While InvokeAI has fantastic in-painting, image-to-image, and ControlNet capabilities, direct instruction-based editing using LLM-guided vision models offers a more conversational, flexible approach to iterative image editing.

Additional Content

No response

[michaelanderson01826-glitch]

Is there already an issue tracking this? I couldn’t find one, but native support for Qwen-VL and instruction-based image editing would be a really nice addition.

…

On Sun, Mar 22, 2026, 7:08 PM Gil Assunção ***@***.***> wrote: *gafda* created an issue (invoke-ai/InvokeAI#8983) <#8983> Is there an existing issue for this? - I have searched the existing issues Contact Details *No response* What should this feature add? This request is to add native support for integrating and running the Qwen series of multimodal models within InvokeAI, specifically targeting the highly capable models from Alibaba Cloud. I'm requesting support for: 1. *Qwen-VL (Vision-Language):* Integration would enable features like advanced image-to-text generation, detailed captioning, visual question-answering, and general image analysis. This could be used to enhance prompt generation. 2. *Qwen Image Edit / Agent:* Support for Qwen's instruction-driven image editing capabilities. This would allow users to perform complex, natural-language-guided image manipulations and stylistic modifications without relying solely on traditional in-painting masks. Alternatives While InvokeAI has fantastic in-painting, image-to-image, and ControlNet capabilities, direct instruction-based editing using LLM-guided vision models offers a more conversational, flexible approach to iterative image editing. Additional Content *No response* — Reply to this email directly, view it on GitHub <#8983?email_source=notifications&email_token=B7HTEEOXUIAIXIMMAJRCIKT4SB57HA5CNFSL4Z3JMQ5C6L3HNF2C22DVMIXUS43TOVSS6NBRGE3TINBWHA2TNJTSMVQXG33OVJZXKYTTMNZGSYTFMSSWK5TFNZ2KYZTPN52GK4S7MNWGSY3L>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/B7HTEELNT43NFIBZSMERGNT4SB57HAVCNFSM6AAAAACW3BHHKSVHI2DSMVQWIX3LMV43ASLTON2WKOZUGEYTONBUGY4DKNQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>