OpenAI Becomes Preferred Model for Microsoft 365 Copilot Amid MAI Cost-Saving Shift
OpenAI will be the preferred model powering Microsoft 365 Copilot, reinforcing ties as Microsoft expands use of in-house MAI models to trim AI costs.
OpenAI said Thursday that its latest model family will serve as the “preferred model” for Microsoft 365 Copilot, signaling the continued integration of OpenAI technology across Microsoft’s productivity suite. The announcement follows recent reporting that Microsoft has been deploying its own MAI models in apps such as Word and Excel as part of broader cost-control measures. The move clarifies that, even as Microsoft diversifies its model mix, OpenAI models remain central to Copilot’s roadmap.
OpenAI’s GPT 5.6 named preferred model
OpenAI launched its new GPT 5.6 model family and identified GPT 5.6 as the preferred model for Microsoft’s 365 Copilot offering. The company said the model will support features across Word, Excel, PowerPoint and Cowork.
OpenAI framed the designation as a continuation of its partnership with Microsoft and emphasized a shared goal of bringing advanced AI to more organizations. The firm did not provide a technical definition of “preferred,” leaving questions about deployment scope and fallback arrangements.
Microsoft’s recent MAI deployment and cost motives
Journalists recently reported that Microsoft is supplementing OpenAI models with its internally developed MAI models to reduce compute expenses. Those in-house models have reportedly been used selectively in productivity apps to manage costs tied to large-scale third-party model usage.
Microsoft has not said it will stop using OpenAI models for Copilot, and the company continues to maintain commercial and technical ties with OpenAI. The combination of in-house and partner models reflects a pragmatic approach to balancing performance, control and operating expenses.
What ‘preferred model’ likely signals operationally
Industry observers say the “preferred model” label is less about exclusivity and more about default routing and quality assurance. In practice, Microsoft may continue to route many enterprise Copilot queries to OpenAI models while using MAI models for lower-cost or specialized tasks.
OpenAI’s statement suggests continued certification, support and optimization for Copilot workloads. That practical alignment would preserve customer experience while allowing Microsoft to leverage cheaper internal models where appropriate.
Analysts weigh partnership stability and commercial strategy
Analysts note that the announcement reduces ambiguity about the Microsoft-OpenAI relationship after a period of mixed signals. Market commentators had speculated that increasing MAI use signaled a strategic distancing; the preferred-model designation counters that narrative without negating Microsoft’s cost-management efforts.
Financial and technical analysts caution that commercial dynamics will continue to evolve as both companies negotiate pricing, model licensing and product differentiation. The extent to which Copilot traffic is divided between OpenAI and MAI will probably depend on cost, latency and performance trade-offs.
Potential impacts on enterprise customers and competition
For enterprise customers, the announcement offers short-term clarity: OpenAI models will remain a supported option for core Copilot features. Customers focused on accuracy and cutting-edge capabilities may favor sessions routed to OpenAI models, while cost-sensitive use cases might rely on MAI deployments.
On the competitive side, Microsoft’s dual approach could influence other cloud and AI vendors to mix proprietary and third-party models. It may also accelerate product segmentation, with premium features tied to higher-tier models and cost-optimized capabilities delivered by internal engines.
OpenAI and Microsoft have repeatedly framed their work as complementary: OpenAI provides leading-model innovation while Microsoft embeds and distributes that technology at scale. The preferred-model announcement reinforces that commercial relationship while acknowledging a new era in which hyperscale providers also push more internal models into production.
OpenAI’s designation as the preferred model for Microsoft 365 Copilot clarifies a recent period of uncertainty but leaves open operational questions about traffic routing and cost allocation. Both companies appear to be pursuing a hybrid strategy that preserves OpenAI’s role in Copilot while allowing Microsoft to pursue internal efficiencies through MAI.
