AuraTracer智迹闻
中文

EVENT DOSSIER

Microsoft released MAI-Image-2.6-Flash: image generation speed doubled and costs significantly reduced.

2026-09-05 14:58 Models 🔥 47.3 heat score
2sources
1days unfolding
47.3heat score
4mentions
SummaryAI generated

On September 4, Microsoft launched its own image generation model, MAI-Image-2.6-Flash, through the Microsoft Foundry platform, and made it available for public preview. Based on the core capabilities of the flagship version MAI-Image-2.6 (such as object replacement and local redrawing), this model significantly improves performance and cost efficiency: image generation speed is doubled, GPU efficiency is increased by 72%, and the price of the text input API is 65% lower than previous versions. Microsoft recommends this model for scenarios that are sensitive to latency, throughput, and cost, while complex editing or high-demand tasks still require the flagship version.

Related eventsRELATED EVENTS
Key entitiesKEY ENTITIES
MAI-Image-2.6-FlashMicrosoftMicrosoft FoundryMustafa Suleyman

Coverage · reports per dayLANGUAGE SPLIT

Entity relations
MAI-Image-2.6-Flash × M…2MAI-Image-2.6-Flash × M…2Microsoft × Microsoft F…2MAI-Image-2.6-Flash × M…1Microsoft × Mustafa Sul…1Microsoft Foundry × Mus…1

Integrated timelineUNIFIED TIMELINE

  1. 2026-09-05

    Microsoft released MAI-Image-2.6-Flash:…

    Microsoft has released the MAI-Image-2.6-Flash self-developed image generation model and made it available for public p…

    2 reports

SignalsSIGNALS

Keyword heat
  • Microsoft2
  • MAI-Image-2.6-Flash2
  • Microsoft Foundry2
  • Mustafa Suleyman1

All reports (2)SOURCES

D DoNews·快讯 zh 2026-09-05 08:00

“Microsoft released MAI-Image-2.6-Flash image generation model”

On September 4, Microsoft launched its self-developed image generation model MAI-Image-2.6-Flash through the Microsoft Foundry platform, making it available for public preview. This model possesses capabilities for text-to-image and image-to-text generation, supporting operations such as object replacement and local redrawing. Compared to the flagship version MAI-Image-2.6, the Flash version significantly optimizes generation speed and cost while maintaining core functions – the generation speed is doubled, and GPU efficiency is increased by 72%. The API is charged per token, with text input costing only $1.75 per million tokens, a 65% reduction from previous versions. Microsoft recommends using this model in scenarios that are sensitive to latency, throughput, and cost.

I IT之家 zh 2026-09-05 14:58

Microsoft released MAI-Image-2.6-Flash: the speed of generating images is twice that of the best AI models in the world.

Microsoft has released the MAI-Image-2.6-Flash self-developed image generation model and made it available for public preview through Microsoft Foundry. The API prices are as follows: text input (1.75 US dollars per million Tokens, approximately 11.8 yuan); image input (2.50 US dollars per million Tokens, approximately 16.8 yuan); image output (19 US dollars per million Tokens, approximately 127.8 yuan). MAI-Image-2.6-Flash shares the core capabilities for text-to-image and image-to-image/image editing with the flagship version MAI-Image-2.6, including object removal/replacement, local repainting, property modification, text updating, etc. Although the Flash version focuses on optimizing generation speed and cost, Microsoft states that it still retains the core generation and editing capabilities. Microsoft recommends using MAI-Image-2… for tasks requiring high-quality final output, complex editing, commercial design, and high requirements for image quality and text rendering.