Open weightsAlibaba (Qwen)/10 trustMetrics-checked

Qwen-Image-Edit

Name: Qwen-Image-Edit
Rating: 7

Apache-2.0 instruction-based image editing — precise text-and-image-to-image edits from the Qwen team.

Visit official page →

About

Qwen-Image-Edit (August 2025) is Alibaba's open instruction-based image editing model: describe the edit in natural language ("change the sky to sunset, keep everything else") and it performs precise text-and-image-to-image generation. Built on the Qwen-Image foundation with strong bilingual text rendering.

Apache 2.0 licensed — commercial use allowed. With 113k Hugging Face downloads it is the most-used open image-editing checkpoint. Part of the broader Qwen-Image family (base text-to-image plus edit variants). Strengths: instruction following, Chinese/English text rendering in edits, and integration with the diffusers ecosystem. For pure generation see Qwen-Image (already in the catalog); this entry covers the editing variant.

Specs

Context window

N/A (image editing model; text instruction + input image)

Pricing

Free — open weights (Apache 2.0), self-hosted, commercial use allowed.

Free quota

Free — download and run the weights yourself (Apache 2.0).

API access

Self-hosted (diffusers; Hugging Face checkpoints).

Weights

Open weights

Released

August 19, 2025

Benchmarks

Provider-claimed (Qwen team): strong instruction-following and text-rendering in edits — treat as provider-claimed. Independent standardized editing benchmarks are limited.

Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.

Use cases

image-editingimage-generationdesignmarketinge-commerce

Trust score — 7/10

/10 trustMetrics-checked

Scored 7/10 (metrics tier). Apache 2.0, the most-used open image-editing checkpoint (113k downloads), and genuinely useful instruction-based editing. Docked three points: (1) quality claims are provider-reported, not independently benchmarked; (2) editing quality lags the best closed systems (GPT Image, Seedream) on complex multi-element edits; (3) the base model's 2025 release means the maintenance signal is aging.

How the trust score is built

Full methodology →

Adoption · stars/downloads, log-scaled vs peers: 7.6/10
Maintenance · recency, releases, issue load: 5.0/10
Openness · license + docs: 9.1/10
Manual review · sanity read + red-flag check: 8.0/10

· HF downloads 113,273 (peer-normalized)
· released 420 days ago
· manual review: most-used open image-editing checkpoint; editing quality lags best closed systems on complex edits; license: Apache-2.0