Microsoft launches MAI-Image-2.5-Pro and MAI-Voice-2-Flash preview, touts up to 89% cost cuts

Microsoft unveiled a public preview of two in‑house AI models, MAI-Image-2.5‑Pro for image generation and MAI-Voice-2‑Flash for speech synthesis. The company highlights substantial operating savings, reporting up to an 84% reduction in GPU costs for PowerPoint and up to an 89% reduction for Dynamics 365 Contact Center compared with earlier models. MAI‑Voice‑2‑Flash is described as twice as fast as its predecessor and 32% cheaper per unit of use. Pricing for the preview is listed at roughly €4.60 per million text tokens, $5 per million text tokens, $8 per million image inputs and $106 per million image outputs, with voice usage priced at around $15 per million characters. The models are available through Microsoft Foundry, underscoring a shift toward AI ownership that reduces reliance on external services and aligns with emerging EU AI‑Act requirements.