I Tested All 4 of Microsoft's New AI Models. Here's the Brutal Truth

The models include MAI-Thinking-1, a reasoning model, and MAI-Image-2.5, an image generation model. The reviewer concluded that the models are "fine, but nothing more," indicating they are competent but not groundbreaking. The review is based on hands-on testing of all four models.

The phrase "fine but nothing more" is a direct quote from the article, which the reviewer uses to summarize the overall impression. The excerpt does not provide specific benchmark scores or detailed comparisons with other models, but the verdict suggests that Microsoft's offerings do not surpass competitors. The names MAI-Thinking-1 and MAI-Image-2.5 follow a numbering scheme that suggests iterative development.

The "fine but nothing more" verdict may influence how enterprises perceive Microsoft's AI capabilities relative to more advanced rivals. Microsoft has not responded to the review.