OpenAI released Images 2.6 on Nov. 4, adding GIF creation and more precise editing controls to its visual tools as creative teams assess where AI fits into production workflows.
Rapid News Brief
Animation joins the editing workflow
The Verge reported that ChatGPT users can create a GIF from a written request or uploaded photograph, change its playback speed and edit individual frames. Those controls extend the task beyond producing a single finished still image.
OpenAI said Images 2.6 is available globally in ChatGPT and Codex across subscription tiers. Developers also get two API models, GPT-Image-2.6 Flare and Sunburst, alongside image markup for specifying edits, SVG export and a complexity setting separate from image quality.
Speed claims need a workflow test
OpenAI says the update generates images 50% faster, but its announcement does not identify the comparison model or testing method. Its separate claim that Flare can produce an image in as little as six seconds applies to the low-quality setting, not every request.
The operational question is whether an acceptable result arrives sooner after revisions and review, not just whether an initial image appears faster. For a creative team evaluating the release, a useful comparison would hold the brief and approval standard constant while tracking editing time, output quality and the cost of rejected versions. Those are evaluation criteria, not demonstrated outcomes of this launch.
The company’s system card says GIF generation is supported in ChatGPT and the API. It also warns that harmful meaning can emerge across frames and that motion-related violations remain difficult to detect. Its adversarial safety tests do not measure how often harmful outputs reach ordinary users.
OpenAI also says it is combining C2PA provenance metadata with Google DeepMind’s SynthID watermarking across ChatGPT, Codex and the API. Those are identification measures, separate from whether an output meets a creative brief.
That distinction matters for brand approval: checking a selected frame is not the same as reviewing the complete animation.
Disclosure: This report was researched and drafted with assistance from OpenAI technology.