I put ChatGPT’s new image model to the test.
Not once, but nine times head-to-head vs. Nano Banana Pro.
And I gotta say: ChatGPT does not measure up here.
The new ChatGPT Images model is not ready for serious business work. Not close.
I’m not saying this to be contrarian. I ran the tests. I have the side-by-side comparisons, and I’ve included them all here.
ChatGPT’s failure modes are consistent enough that we can learn from them how OpenAI is solving the image problem, and we can infer a fair bit of how that differs from Google and Nano Banana.
Here’s what’s inside:
The fundamental difference in how these models generate diagrams—ten takeaways around how the models build images differently, and why that matters for production systems.
Nine side-by-side visual comparisons across ARR revenue bridges, funnel diagnostics, Venn diagrams, opportunity solution trees, fictional maps, and image editing tasks—with my assessment of what worked and what failed.
82 prompts across three categories (advertising, business artifacts, education) that I built to extract what image models actually do well. Structured interviews, not generic templates.
20 Board Level Business Slide Creation Prompts
Financial: ARR revenue bridge, P&L waterfall, unit economics dashboard, pricing tier comparison
Metrics: Funnel diagnostic, cohort retention heatmap, engagement overview, North Star tracker
Strategy: Competitive positioning map, market landscape, SWOT analysis, strategy-on-a-page
Product: Opportunity solution tree, roadmap timeline, prioritization matrix, feature comparison grid
Planning: OKR cascade, project timeline, RACI matrix, risk assessment
17 Advertising Image Creation Prompts
Food & Beverage
SaaS / Tech Products
Real Estate Listings
Financial Services
Healthcare & Wellness
Automotive
Travel & Hospitality
Fashion & Apparel
Consumer Electronics
B2B Professional Services
Education & EdTech
Non-Profit Campaigns
Event Promotion
Recruitment & Employer Branding
Retail & E-commerce
Luxury Goods
Local Small Business
45 Education Worksheet Image Creation Prompts
Mathematics (8): Arithmetic practice, word problems, fraction models, geometry exploration, multiplication tables, number lines, place value, fact fluency
Reading & Language Arts (8): Phonics, comprehension organizers, vocabulary builders, creative writing, sight words, story sequencing, grammar, spelling patterns
Science (6): Labeled diagrams, experiment logs, classification sorts, nature journals, scientific method templates, life cycles
Social Studies (5): Map skills, timelines, civics organizers, geography labeling, historical comparisons
Early Childhood (6): Letter tracing, counting, shapes, patterns, colors, fine motor
Specialized Formats (12): Word searches, crosswords, bingo, cut-and-paste, foldables, matching, fill-in-blank, multiple choice, true/false, sequencing cards, Venn diagrams, KWL charts
A quick reference guide so you know which prompts to grab for board prep vs. quarterly planning vs. launch readiness.
Why so many prompts? Why so many categories? Well, honestly my response to a disappointing launch is to build. So as soon as I saw ChatGPT’s image launch was disappointing, my instinct was to turn around and build prompts for Nano Banana Pro. I wanted to show range, so I thought going across education, advertising, and board slides would be a nice roundup.
The other reason to throw these all in here is simple: I write a lot, and I think sometimes people lose track of the stuff I write. Having a really fat prompt collection like this makes it easy for folks to save and come back to these collections because it’s all in one handy place.
With that, let’s dig in, get those prompts going, and see the side-by-side comparison!
Listen to this episode with a 7-day free trial
Subscribe to Nate’s Substack to listen to this post and get 7 days of free access to the full post archives.













