OpenAI Updates Data Export Features: What's New in Your ChatGPT Archive

OpenAI has released an update to ChatGPT's data export functionality, addressing several limitations that users and developers had flagged with the previous export format. The updated export now includes richer metadata fields, improved JSON structure for programmatic processing, and significantly faster archive generation for accounts with large conversation histories.

Key changes in the updated export format include a new message metadata field that captures model version, temperature settings, and tool usage per message — information previously absent from exports. The conversation title generation algorithm has also been improved, making auto-generated titles substantially more descriptive and useful for identifying specific conversations in exported archives.

Developers building tools on top of ChatGPT export data have welcomed the improved structure, noting that the previous format's inconsistent handling of multi-modal content (conversations containing images, file uploads, or code interpreter outputs) made parsing complex. The updated format introduces a consistent content_parts array that represents all content types uniformly.

Archive generation time for large accounts has dropped from multiple hours to under 30 minutes in most cases, with OpenAI attributing the improvement to infrastructure changes in their data pipeline. Export requests are now processed on a dedicated infrastructure path rather than competing with production traffic.