Microsoft is shifting Copilot and Office products to rely more on its own cheaper in-house AI models instead of those from OpenAI and Anthropic, aiming to reduce its third-party AI spending over time. For end users, this may mean paying the same subscription price for less capable AI; Microsoft is also considering usage-based pricing where cheaper in-house models are the default and premium third-party models carry an extra surcharge.
Summaries like this, in your inbox every morning.
Sign up free →What happened
Microsoft is replacing AI models from OpenAI and Anthropic with its own in-house models across Copilot products including Excel and Outlook. The in-house models are already processing tens of thousands of requests per week, though they still handle only a small fraction of total requests. Microsoft also unveiled seven new AI models at Build, including MAI-Thinking 1, its first reasoning model.
Why it matters
Copilot and Office customers may end up paying the same price for weaker AI as Microsoft cuts costs. Microsoft's head of AI stated in June the goal is to "reduce and ultimately eliminate" spending on Anthropic. CEO Satya Nadella has also hinted that pricing could shift to usage-based models, potentially making cheaper in-house models the default with third-party models available as premium add-ons—passing costs to customers as a surcharge.
What to watch
Microsoft claims its in-house models are trained on clean, commercially licensed data, but the technical paper shows it used Common Crawl, a freely accessible web dataset whose use for AI training is not legally settled. On benchmarks, MAI-Thinking 1 trailed OpenAI and Anthropic models by a wide margin and landed roughly on par with Deepseek V3.2, despite Microsoft's claims it could match Sonnet 4.6 and Opus 4.6 in coding.
Microsoft's shift toward in-house AI models represents a direct cost-containment strategy. The company's head of AI explicitly acknowledged in June that paying third-party companies like Anthropic is expensive and the goal is to eliminate that expense. This move sits in tension with Microsoft's recent public stance: the company has argued that vendor lock-in with OpenAI and Anthropic is problematic and that it wants to be a platform-neutral alternative. However, by making its own models the default and pricing third-party models as premium add-ons, Microsoft would effectively lock customers into its own AI offerings while shifting costs to those who want more capable competitors' models.
The technical performance gap is notable. Microsoft publicly claimed its new MAI-Thinking 1 reasoning model could match Anthropic's Sonnet 4.6 and Opus 4.6 based on human evaluations, but released benchmarks told a different story—the model trailed both OpenAI and Anthropic significantly and performed roughly on par with Deepseek V3.2. This suggests either the human evaluation criteria differ markedly from standard benchmarks, or the benchmarks reveal weaker real-world capability than Microsoft's marketing claims. For business customers accustomed to high-capability AI assistants, the transition to cheaper in-house models as the default could represent a visible quality trade-off for the same price.
No comments yet. Be the first to share your thoughts!
Log in to join the discussion
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
1 minute a day. The AI essentials.
200+ sources · Email / LINE / Slack
