Mira has launched the Verify system, which is used to mark the judgment of the accuracy of AI-generated content, and if the output result is not confirmed by multiple other models, then it will be marked as an uncertain result, which can be used with the API interface as a whole. After looking at the effect, it's really a useful feature. For example, in the following one, the error of the red part is actually very small. Because the truth is that Raoul Pal quit at the age of 36, not at the age of 38. If you're not familiar with this guy, it's still hard to find out this error. This kind of verification, at least for now, is very valuable from the perspective of practical application.