Dynamic Vision Beating AI News Flash: Microsoft AI has launched MAI-Image-2.6-Flash and made it available for public preview on the Microsoft Foundry. It is an accelerated version of the flagship model MAI-Image-2.6, which also supports text-to-image, image editing, multi-image reference, online information retrieval, and automatic aspect ratio selection, with a focus on further reducing generation latency and costs.

Speed is the key highlight of this release. The rendering speed of MAI-Image-2.6-Flash is 2.8 times faster than GPT-Image-2 Medium, with a high GPU utilization efficiency of 72%.

However, it is not simply about sacrificing quality for speed. In the current Artificial Analysis image editing leaderboard, MAI-Image-2.6-Flash ranks third with 1311 Elo, slightly higher than GPT Image 2 high with 1309 Elo; in text-to-image, it ranks eighth with 1297 Elo. The flagship MAI-Image-2.6 performs even better, currently ranking first in image editing and second in text-to-image.

On the Microsoft Foundry, the image output price of MAI-Image-2.6-Flash is $19 per million tokens, while the flagship 2.6 is priced at $38. Based on representative image calculations by Artificial Analysis, the Flash model costs approximately $19.5 per thousand images, compared to around $211 per thousand images for GPT Image 2 high.