OpenAI launches GPT Image 1.5 and rolls back ChatGPT router for free users
OpenAI unveiled its new flagship image generator GPT Image 1.5 with faster processing and improved editing capabilities, while reverting ChatGPT's automated model router for free and budget-tier subscribers.
December 16, 2025 - 06:56 PM ET • 2 min read
OpenAI announced Tuesday the launch of GPT Image 1.5, described as its new flagship image generation model, available immediately to all users. The model offers improved instruction-following capabilities, faster processing speeds up to four times quicker than previous versions, and enhanced photo editing functionality.
The company said the new model will better "match your intent," particularly when modifying existing photographs. Specific improvements include more useful photo edits, more believable clothing and hairstyle try-ons, alongside stylistic filters and conceptual transformations that retain the essence of the original image, according to OpenAI's blog post.
The image generation tool will be accompanied by a redesigned interface featuring a new dedicated Images tab in the ChatGPT sidebar. This tab will include preset filters and trending prompts and ideas to guide users in generating images.
In a separate development, OpenAI rolled back a major change to its ChatGPT model router system for users on its Free and $5-a-month Go tiers. The model router, which launched four months ago as an automated system to direct complex user questions to more advanced reasoning models, will no longer operate by default for these user groups.
Free and Go tier users will now default to GPT-5.2 Instant, OpenAI's fastest and cheapest-to-serve model version. Users on these tiers will retain the ability to access reasoning models, but must select them manually rather than having the system automatically route their queries.
The model router was introduced as part of OpenAI's effort to unify the user experience following the debut of GPT-5. The system was designed to analyze user questions and determine whether to route them to fast-responding, inexpensive AI models or slower, more expensive reasoning models based on query complexity.
OpenAI announced the rollback on a low-profile blog that tracks product changes. The company did not provide a public statement explaining the decision to revert the feature for free and budget-tier users.