The New ChatGPT Images: What Actually Changed (And How to Use It) [2025]

OpenAI's GPT Image 1.5 is 4x faster, finally handles text properly, and won't destroy your edits. Here's what matters for businesses.

If you've used AI image generation before, you know the frustrations: slow generation times, text that looks like it was written by someone having a stroke, and the infuriating habit of models completely reimagining your image when you ask for a simple edit.

GPT Image 1.5 addresses all three.

What's Actually New?

OpenAI announced this update on December 16, 2025, positioning it as their response to Google's Gemini 3 and the viral Nano Banana Pro image generator that's been eating into ChatGPT's market share. The launch reportedly accelerated from an early January release date after Google's models started topping the LMArena leaderboard.

For businesses using AI-generated images—whether for marketing, product catalogs, or internal content—this is the first image model that might actually fit into professional workflows without constant supervision.

Key Features at a Glance

• 4x faster generation compared to previous models
• Edit preservation — no more complete re-renders when you request changes
• Better text rendering — denser, smaller, more accurate
• New Images interface with preset filters and trending prompts
• 20% cheaper API pricing — $0.01 to $0.17 per square image
• Available now for all ChatGPT users; API access as gpt-image-1.5

The Three Things That Actually Matter

1. Edits That Don't Destroy Your Image

This is the headline improvement. Previous AI image models had a frustrating habit: ask for a small change, get a completely different image.

When users upload a photo and ask for changes—say, swapping a shirt or adding a hat—the model now adheres to the intent reliably, preserving crucial elements like lighting, composition, and even facial likeness across subsequent edits.

What this means in practice:

• Change someone's outfit without changing their face
• Adjust lighting without losing composition
• Add or remove elements while keeping the rest intact
• Iterate on designs without starting over

The model excels at different types of editing—including adding, subtracting, combining, blending, and transposing—so you get the changes you want without losing what makes the image special.

For product photography, marketing materials, or any workflow requiring iteration, this is a fundamental shift from "AI as a slot machine" to "AI as an actual editing tool."

2. Text That's Actually Readable

Text rendering has been the Achilles' heel of AI image generators. Letters would merge, words would scramble, and anything beyond a few large characters looked like abstract art.

The new ChatGPT Images is capable of handling denser and smaller text—a significant improvement for anyone creating images that need to include words (which is most business use cases).

This opens up practical applications that were previously unreliable:

• Social media graphics with quotes or statistics
• Product mockups with labels
• Presentation slides with embedded visuals
• Marketing banners with actual copy

Still not perfect for dense paragraphs or legal fine print, but usable for most marketing and communication needs.

3. Speed That Doesn't Kill Your Workflow

The new version of ChatGPT Images is designed to make and edit images more precisely, as well as to spit out pictures as much as four times faster than its previous AI image-generation model.

If you've sat waiting 30-60 seconds per image during peak times, you understand why this matters.

Faster generation means:

• Quicker iteration cycles
• More viable for real-time use cases
• Less friction when exploring creative directions
• Actually usable during client calls or presentations

The New Images Experience in ChatGPT

Beyond the model improvements, OpenAI added a dedicated Images section to ChatGPT's interface.

ChatGPT images will also now be accessible via a dedicated entry point in the ChatGPT sidebar that works "more like a creative studio."

What's Included

Preset filters: Quick style applications without crafting prompts

Trending prompts: Inspiration from what's working for others

Image library: All generated images saved automatically for later access

Editing screens: Dedicated interface for refining images

"The new image viewing and editing screens make it easier to create images that match your vision or get inspiration from trending prompts and preset filters."

This is OpenAI positioning ChatGPT as a creative tool rather than just a chatbot—part of their broader push to make it an "everything app."

API Pricing (GPT Image 1.5)

For developers and businesses building on the API:

The API version, GPT Image 1.5, is now 20% cheaper for image inputs and outputs.

Token-Based Pricing Breakdown

• Text input: $5 per 1M tokens
• Image input: $10 per 1M tokens
• Image output: $40 per 1M tokens

Non-square images (portrait or landscape) cost proportionally more based on pixel count.

Practical Cost Examples

• 1,000 medium-quality social media images: ~$40
• 100 high-quality product shots: ~$17
• 10,000 low-quality thumbnails: ~$100

For most business applications, the cost per image is negligible compared to the time saved versus manual design or traditional stock photography.

How Businesses Are Using This

Companies like Wix are already leveraging the model's improved consistency for brand work, noting that the ability to preserve logos and key visuals across edits is essential for generating product catalogs and marketing materials at scale.

E-commerce Product Images

Upload a product photo, then generate variations:

• Different backgrounds
• Lifestyle contexts
• Color variations
• Seasonal themes

The edit consistency means your product stays recognizable across variations.

Marketing Content at Scale

Generate consistent visual content for:

• Social media campaigns
• Email headers
• Blog post images
• Ad creative testing

The improved text rendering makes it viable for content that includes copy.

Brand Asset Creation

• Product mockups
• Presentation visuals
• Internal communications
• Training materials

The faster generation and better editing make iteration practical rather than painful.

Integrating Image Generation Into Your Workflows

Here's where no-code automation becomes powerful.

With Arahi AI, you can connect GPT Image 1.5 to your existing business tools and automate image generation as part of larger workflows:

Example: Automated Social Media Content

1. New blog post publishes → triggers workflow
2. AI extracts key points and suggested visuals
3. GPT Image 1.5 generates branded social graphics
4. Images route to approval queue or post directly
5. Variations generated for different platforms

Example: Product Catalog Updates

1. New product added to inventory system
2. Product photos processed through image generation
3. Lifestyle shots and variations created automatically
4. Images pushed to e-commerce platform
5. Marketing team notified for review

Example: Personalized Marketing

1. Customer segment identified
2. Relevant visual style selected based on preferences
3. Personalized images generated
4. Content pushed to email or ad platform
5. Performance tracked and fed back for optimization

The key is that image generation becomes one step in a larger automated process rather than a standalone task requiring manual intervention.

What GPT Image 1.5 Still Can't Do

OpenAI is upfront about limitations:

OpenAI acknowledges that results remain "imperfect" in areas like scientific accuracy and rendering multiple small faces.

Current Limitations

• Complex scenes with many people: Faces can still get weird
• Scientific/technical accuracy: Don't use it for diagrams requiring precision
• Very dense text: Headlines yes, paragraphs no
• Consistent characters across images: Better, but not reliable enough for comic strips or sequential storytelling
• Photorealistic humans in specific poses: Can be uncanny

Where It Excels

• Product photography and mockups
• Abstract and stylized graphics
• Marketing visuals with text overlays
• Conceptual illustrations
• Background and environmental images
• Quick iteration and ideation

GPT Image 1.5 vs. Previous Options

The jump from DALL-E 3 to GPT Image 1.5 is substantial enough that workflows built around the older model should be re-evaluated.

Getting Started

For ChatGPT Users

GPT Image 1.5 is rolling out now to all users (Free, Plus, Pro, Team). Look for the new Images section in the sidebar, or simply ask ChatGPT to generate images as you normally would.

For Developers

Access via the API as gpt-image-1.5. The model uses the same endpoint structure as previous image models.

# Example API call
curl https://api.openai.com/v1/images/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -d '{
    "model": "gpt-image-1.5",
    "prompt": "A professional product photo of wireless earbuds on a marble surface",
    "n": 1,
    "size": "1024x1024",
    "quality": "high"
  }'

For Business Automation

Platforms like Arahi AI let you integrate image generation into workflows without coding. Connect your CRM, e-commerce platform, or content management system, define triggers and conditions, and let AI handle the image creation as part of your existing processes.

The Bottom Line

GPT Image 1.5 represents a meaningful shift from "AI image generation is a fun toy" to "AI image generation fits into business workflows."

The combination of faster generation, consistent editing, and improved text rendering addresses the three biggest practical barriers to business adoption. The 20% price reduction on API access doesn't hurt either.

For businesses already using AI in their workflows, this is worth integrating. For those still on the sidelines, the edit consistency alone makes this worth testing for product photography, marketing content, or any visual workflow that currently requires manual iteration.

Want to automate AI image generation workflows? Arahi AI connects GPT Image 1.5 to 2,800+ apps and builds intelligent automations—no code required.