# Microsoft unveils new in-house AI models claiming up to 89% cost savings versus OpenAI

With MAI-Image-2.5-Pro and MAI-Voice-2-Flash, Microsoft strengthens its generative AI infrastructure.

- **Source:** [VentureBeat (feeds.feedburner.com)](https://brifing.si/source/c51474dc-6b09-49dd-aaca-085f4ed00c84)
- **Published:** 2026-07-23
- **Topic:** [Data & AI](https://brifing.si/c/data-ai)
- **Original article:** <https://venturebeat.com/infrastructure/microsoft-launches-new-in-house-ai-models-it-says-cut-costs-up-to-89-versus-openai>
- **On Brifing:** https://brifing.si/s/9243449c-1b40-4eff-a09c-bb3540ee7d38

## Microsoft's AI breakthrough

Microsoft introduced two new AI models, MAI-Image-2.5-Pro and MAI-Voice-2-Flash, marking a significant step in reducing reliance on OpenAI. These specialized models for image generation and speech processing are being deployed across various industries.

## Revolution in image generation

MAI-Image-2.5-Pro excels in detailed editing and text rendering, ranking second on Arena's leaderboard. The model has received positive feedback from the advertising industry, with WPP's CCO calling it a leap forward for GenMedia tools.

## Efficiency in voice technology

Optimized for call centers and real-time applications, MAI-Voice-2-Flash offers 32% cost reduction and twice the speed of its predecessor. Major companies like T-Mobile and EasyJet have already adopted this technology.

## Product integrations

The new models power Microsoft products including Bing, PowerPoint, and GitHub Copilot. Bing Image Creator now runs entirely on MAI-Image-2.5, while PowerPoint reports 84% lower GPU costs compared to OpenAI's models.

## Healthcare transformation

Dragon Copilot, serving 170,000 medical providers in 58 languages, has reduced transcription errors by 50% with MAI-Transcribe-1.5. This improvement significantly enhances the accuracy of clinical documentation.

## Small model success in Excel

MAI-Code-1-Flash matches GPT-5.6's performance on common Excel tasks while running on older GPUs. User retention rates are 11% higher compared to competing models, demonstrating its effectiveness.

## Cost advantages

Microsoft reports up to 89% GPU cost savings with its in-house models, with the highest savings in Dynamics 365 Contact Center. OneDrive has seen 26% higher save rates with the new implementation.

## Technology strategy

The company's 'hill-climbing machine' methodology optimizes smaller models through integrated data-model-product cycles. This approach gives Microsoft a competitive edge in AI development.

## Industry implications

Microsoft's move sets an example for companies developing proprietary AI infrastructure. Cost efficiency and specialized models are likely to shape the future of the AI industry.

## Sources covering this story

- https://venturebeat.com/infrastructure/microsoft-launches-new-in-house-ai-models-it-says-cut-costs-up-to-89-versus-openai — 

---

_Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._