AI Image Editor
Version 1.2.0
November 14, 2026
Added
- Added the GPT Image 2.6 models
gpt-image-2.6-flareandgpt-image-2.6-sunburst, with the same 2K tier andxhigh/maxquality levels as 2.5 - Added a
complexitysetting for the GPT Image 2.6 models. When it's not set, edits with "Precise edits" on uselowcomplexity - Added the
nano-banana-2.1model alias (gemini-nano-banana-2.1), with resolution tiers from 512 to 4K and thinking level support - Added the FLUX 3 Image models
flux-3-imageandflux-3-image-klein, and agroundingsetting forflux-3-image(off by default)
Changed
- Changed the default OpenAI model to
gpt-image-2.6-flare, and the default FLUX model toflux-3-image. SetdriverConfig.<handle>.defaultModelto keep the old ones - Changed the OpenAI analysis model to
gpt-6.1-terra - Changed the retired
gpt-image-1,gpt-image-1.5andgpt-image-1-minimodels (shut down by OpenAI on December 1, 2026) to be swapped forgpt-image-2.6-flare, with a deprecation warning logged. "Precise edits" no longer setsinput_fidelity, since no remaining model accepts it - Changed the retired
grok-imagine-image-qualitymodel (retired by xAI on November 2, 2026) to be swapped forgrok-imagine-image-2.0, with a deprecation warning logged
Version 1.1.0
September 28, 2026
Added
- Added the GPT Image 2.5 models
gpt-image-2.5-flareandgpt-image-2.5-sunburst, which reach 2K and support thexhighandmaxquality levels - Added the
grok-imagine-image-2.0model, and aqualitysetting for it - Added the
flux-2-maxandflux-2-klein-9bmodels
Changed
- Changed the default models to
gpt-image-2.5-flare,grok-imagine-image-2.0andflux-2-max. Grok and FLUX cost more per image than the previous defaults, so setdriverConfig.<handle>.defaultModelto keep the old ones - Changed the OpenAI driver to expose a
2Kresolution tier on the 2.5 models, where accepting a result now regenerates at the final resolution - Changed the analysis models to
gemini-3.8-flash,gpt-5.4-miniandgrok-4.7
Version 1.0.0
August 1, 2026
Added
- Initial public release