Hey HN! I built GPT Image 1.5 (https://gpt-image-15.com), a web interface for
OpenAI's latest image generation model released Dec 16, 2025.
**What it does:**
- Text-to-image generation with precise instruction following
- Targeted image editing (e.g., "change shirt color" without touching
face/background)
- Superior text rendering in images (infographics, UI mockups, marketing
materials)
- 4x faster than previous models (~2min for complex prompts, seconds for simple
ones)
**Why I built this:**
OpenAI released the API but the ChatGPT UI has limitations for power users. I
needed:
- Batch generation workflows
- Style consistency across multiple images
- API integration for product mockups
- Faster iteration cycles than ChatGPT's interface
**Tech stack:**
- Next.js 16 (App Router, React Server Components, Turbopack)
- React 19 with React Compiler
- Tailwind CSS v4
- Better Auth + Drizzle ORM (PostgreSQL)
- next-intl for i18n
**Benchmarks:**
- LMArena ranking: #1 for text-to-image (1264) and editing (1409)
- Beats Google Nano Banana Pro in instruction following
- 95% semantic matching between prompt and output
**Honest limitations:**
- Complex prompts still take ~2 minutes (vs seconds for simple ones)
- Text rendering improved but not perfect in extreme cases
- Inherits OpenAI's content policy restrictions
I'd love feedback on the UI/UX, especially from designers and developers using
AI image gen in production workflows.
Live demo available at the homepage generator - no signup needed for the first
2 images.
I can totally relate—I’m also a web dev and spend all day in front of screens. Lately I’ve been feeling really stuck and weighed down, and honestly I’m not sure how to start changing things. Reading your story gives me a little hope, though.
One thing I’ve noticed when comparing these models is that “quality” and “realism” don’t always move together.
Some models are very strong at sharp details and localized edits, but they can break global lighting consistency — shadows, reflections, or overall scene illumination drift in subtle ways. GPT-Image seems to trade a bit of micro-detail for better global coherence, especially in lighting, which makes composites feel more believable even if they’re not pixel-perfect.
It’s hard to capture this in benchmarks, but for real-world editing workflows it ends up mattering more than I initially expected.
