Qwen-Image-2.1 tops open-weights image leaderboards, ranking #18 overall on T2I and editing
ArtificialAnlys · x · 2026-10-02
Artificial Analysis evaluated Alibaba's Qwen-Image-2.1, released with open weights on Sep 20, and found it the #1 open-weights model on both AA-Image leaderboards.
- 7B visual generation component handling text-to-image and editing in one model, with native 2K output and RGBA transparency
- Ranks #18 overall on AA-Image-T2I v2.0 (up from #72 for Qwen Image 2.0) and #18 on Image Editing (up from #58), leading open weights models ahead of Ideogram 4.0 (Quality) and HunyuanImage 3.0 Instruct
- Weights under Qwen Research License (commercial use requires separate license); day-0 support in Diffusers, ComfyUI, vLLM-Omni, SGLang and LightX2V
- Evaluated via local hosting since no major inference providers hosted it at launch
- Best editing use cases: Productivity & Knowledge Work, Retail & Ecommerce, Animation & Gaming; closes the gap to frontier on all ten use cases vs 2.0
More from Models
- Google clarifies Gemini TTS voice policy: custom voices only removed after one year of non-use — AI_Andrew · 2026-10-02
- Behavioural evals 'escaped containment' and cheated on their tests, says exec — gabriel1 · 2026-10-02
- KOL says Opus 5.5 is the first model to genuinely surprise him with its taste — Hesamation · 2026-10-02
- OpenAI clones TypeSafe's Jev just 14 days after debut, eroding its moat — JnBrymn · 2026-10-02
- Grok now answers questions directly inside X group chats via XChat — Baconbrix · 2026-10-02
- Daniel Bourke confirms SigLIP nuanced-text score example was real, from late 2024 — mrdbourke · 2026-10-02