Mona-Lisa-1: Prompts, GPT Image 2.5 Comparison and Release Status
Mona-lisa-1 is a reported Arena image-model alias. Review source evidence, prompt ideas, comparisons with GPT Image 2.5 and what remains unconfirmed.
What is mona-lisa-1?
It is an anonymous or pre-release identifier reported after Arena battles. The observation can be documented without claiming it is GPT Image 2.5 or GPT Image 3.
Mona-lisa-1 vs GPT Image 2.5
Early same-prompt reports focus on more natural skin, less plastic texture, improved scene detail and stronger instruction following. These are test hypotheses, not a model card.
What the newer luna-lisa-alpha reports change
The reported move from mona-lisa-1 to luna-lisa-alpha suggests Mona was a checkpoint in an active test sequence, not necessarily a launch candidate. Newer samples shift the test focus from attractive single images to stability, grain, fine-detail patterns, text, and character consistency.
Prompts to prepare before release
Reusable prompts should test real production jobs rather than one attractive portrait.
Related GPT image resources
Luna-lisa-alpha
Review the newer reported checkpoint alias.
GPT Image 2.5 tracker
Follow the incremental-version theory.
GPT Image 3 comparison plan
See the controlled benchmark framework.
Dated Arena observation
Read EvoLink's Aug 11 record of a revealed mona-lisa-1 battle.
OpenAI provenance guide
See why provenance can establish tool origin but not a model name.
Frequently asked questions
Is Mona-lisa-1 an official OpenAI model?
No official OpenAI product or API documentation uses that name.
Can I generate with Mona-lisa-1 here?
No verified endpoint is available. GPT Image 2.5 currently uses GPT Image 2.5.
Is Mona-lisa-1 the same as luna-lisa-alpha?
They are different reported aliases. A relationship between them has not been officially documented.
Test the same prompts on today's model
Save comparison-ready prompts now, generate with GPT Image 2.5, and reuse them when a verified successor launches.