OpenAI has deployed ChatGPT Images 2.6 across ChatGPT and Codex, adding GIF creation and reference persistence features for users now.

The update intends to streamline multimedia outputs and maintain context across sessions for more consistent creative workflows and prompt reuse.

What’s new

ChatGPT Images 2.6 introduces a native GIF generation capability that converts user prompts and image references into short animated outputs.

The model also remembers user-supplied references across prompts, enabling reuse of style and composition cues without reuploading images repeatedly anymore.

Integration with Codex

OpenAI has integrated the 2.6 image features into Codex, allowing developers to generate and manipulate animated visual assets programmatically via APIs.

Report on GPT-6.2

Separately, The Information reports that GPT-6.2 may arrive on Friday in three configurations: Astra, Sol, and a dedicated cybersecurity edition.

According to the report, Astra is positioned as the high-performance variant while Sol aims to be the cost-optimized option for broader deployment.

The cybersecurity-focused instance is described by the outlet as targeted at threat detection and secure code generation use cases for enterprise customers.

OpenAI has not provided an official schedule for GPT-6.2, and the report frames the timing as tentative pending company confirmation.

Users and developers should look for official statements from OpenAI for release details and documentation prior to any production deployment decisions.