OpenAI's New Image Generation Model: 'Code Red' Warpath Continues

OpenAI Launches GPT Image 1.5 for Enhanced Image Generation

A new iteration of ChatGPT Images is being deployed by OpenAI, promising improvements in adherence to instructions, more refined editing capabilities, and a significant boost to image creation speed – up to four times faster.

This latest model, designated GPT Image 1.5, is now available to all ChatGPT subscribers and through the API, beginning on Tuesday.

Response to Competitive Pressure

The release represents OpenAI’s latest move in the ongoing competition with Google’s Gemini. This follows a “code red” situation declared internally by OpenAI CEO Sam Altman last month, as detailed in a leaked memo.

The memo outlined strategies for OpenAI to reclaim its leadership position in the AI landscape, after Google began gaining market share with the introduction of Gemini 3 and Nano Banana Pro.

Both of these Google offerings have achieved top rankings on the LMArena leaderboard across various performance metrics.

Following the Launch of GPT-5.2

Even after OpenAI’s response last week with the launch of GPT-5.2, positioned as its most advanced model for both developers and general users, Google has maintained its leading position.

Originally planned for an early January release, OpenAI accelerated the rollout of its new image generator with this week’s announcement.

The previous image model release, GPT Image 1, occurred in April.

Advanced Features of GPT Image 1.5

GPT Image 1.5 arrives at a time when image and video generation technologies are maturing, moving beyond experimental stages to offer more robust, production-level functionality.

Similar to Nano Banana Pro, ChatGPT Image now includes post-production features, enabling more precise control over edits while preserving visual consistency.

This includes maintaining aspects like facial features, lighting conditions, overall composition, and color palettes across multiple revisions.

Addressing Iteration Challenges

A common weakness of many Generative AI image tools is their difficulty with iterative refinement.

When asked to make specific adjustments, such as altering a facial expression or adjusting the lighting, models often completely recreate the image, resulting in inconsistencies.

This update aims to overcome this limitation.

Improved User Interface

Beyond new features, ChatGPT images are now accessible through a dedicated entry point within the ChatGPT sidebar.

This new interface functions “more like a creative studio,” according to Fidji Simo, OpenAI’s CEO of applications, in a blog post published Tuesday.

“The updated image viewing and editing screens simplify the process of creating images that align with your vision, or finding inspiration from popular prompts and pre-set filters,” Simo explained.

Enhanced Visual Integration

In addition to the new image generator, OpenAI is introducing enhancements to improve the overall ChatGPT experience with more visual elements.

The company plans to incorporate more visuals into search results, clearly indicating their sources.

This could prove beneficial for tasks such as unit conversions or checking sports scores, as noted by Simo.

“When you are in the process of creation, you should be able to visualize and refine the output.

When images convey information more effectively than text, ChatGPT should incorporate them.

And when a quick answer or the next step resides in another tool, it should be readily available.”

Simo concluded, “Our goal is to continually reduce the gap between your ideas and your ability to bring them to life.”