OpenAI Tests Two Next-Gen Image Models
OpenAI appears to be testing two unreleased image-model checkpoints, “mona-lisa-1” and “luna-lisa-alpha,” reportedly corresponding to a larger GPT Image 2.5 and a smaller model based on GPT Luna. The models are not officially announced, named, or available to developers.
This looks more like credible pre-release testing than a launch, but the lack of official documentation makes every capability claim provisional.
- –“Mona Lisa” reportedly delivers a modest improvement over GPT Image 2, suggesting an iterative upgrade rather than a dramatic leap.
- –“Luna Lisa Alpha” could give OpenAI a cheaper, faster option for high-volume image generation and editing.
- –Unofficial aliases appearing in evaluation arenas often precede public releases, but they do not guarantee shipping timelines or final branding.
- –Developers should wait for API documentation, pricing, rate limits, and reproducible evaluations before planning integrations.
- –OpenAI’s apparent A/B testing underscores how quickly image models are moving toward frequent checkpoint-level iteration.
DISCOVERED
1h ago
2026-08-22
PUBLISHED
2h ago
2026-08-22
RELEVANCE
AUTHOR
mark_k