Whether you're generating assets in the dashboard, using the API, or building out campaigns in the creative suite, you now have access to our most powerful image model yet.
What’s new at a glance?
- Photorealistic details and improved textures
- Incredible text rendering accuracy (yes, it can spell!)
- 30% faster generation times
- Flawless understanding of complex, multi-subject prompts
- Supports a wider range of aspect ratios compared to GPT 1.5.
Pro Tip: You can now dictate not just what the text says, but exactly how it looks and interacts with the environment! Try specifying materials and typography to see what it can really do. Try this prompt: The word "PENCIL AI" carved out of raw, glittering amethyst crystal, illuminated by dramatic studio lighting.
Quick note: If you don't immediately see GPT Image 2 in your model drop-down list, it just needs to be toggled on! Head over to Settings -> AI Governance -> Image Generation Models to enable Image Generation Models for your workspace.
Note: No changes in cost or terms
