OpenAI's Anonymous Image Model mona-lisa-1 Debuts on LM Arena, Surpassing GPT Image 2 in Realism
OpenAI released a new image generation model codenamed mona-lisa-1 on August 10, 2026, debuting on the LM Arena blind test leaderboard. Compared with GPT Image 2, launched four months earlier, the new model produces more realistic images with noticeably reduced plastic-like rendering. It demonstrated strong performance on complex tasks including webpage user interfaces, anatomical diagrams, and handwriting imitation. Date-related tests suggest training was completed in 2025, with a knowledge cutoff of December 2025, matching GPT Image 2. The anonymous version lacks internet access, limiting its handling of recent news. OpenAI has not announced an official release date.
OpenAI released a new image generation model codenamed mona-lisa-1 on August 10, 2026, which has officially appeared on the LM Arena blind test leaderboard. Compared with GPT Image 2, released four months earlier, the new model shows improved image generation quality, with noticeably reduced plastic-like rendering and enhanced realism. In testing, mona-lisa-1 demonstrated high levels of completion in generating complex content such as webpage user interfaces, anatomical diagrams of human organs, and handwriting imitation.
Through five consecutive date-related tests, developers found that mona-lisa-1's training was completed in 2025, with its knowledge cutoff consistent with GPT Image 2 at December 2025. The model failed to process the latest news effectively, in part because the anonymous version does not have internet access enabled.
OpenAI has not yet announced a release date for the official version, but the model's appearance indicates that AI image generation technology continues to evolve.
Why this event matters
The event has a measured impact on 1 industry. The strongest current signal is positive for Artificial Intelligence, with intensity 60/100 and 80% confidence over a short term horizon.
Impact figures are analytical estimates that combine direction, intensity, confidence and event importance. They are not investment advice.