Microsoft's new MAI-Thinking-1 model has an impressive 35 billion parameters, positioning itself as a cost-effective alternative to competitors like GPT and Claude.
During the Build 2026 conference in June, Microsoft showcased seven new in-house models, with MAI-Thinking-1 leading the charge. The company claims these models can achieve cost efficiencies up to ten times better than those of rival models in enterprise applications. This means businesses can expect substantial savings without sacrificing performance, particularly in coding tasks.
For Microsoft, which has a substantial stake in OpenAI valued at around $135 billion, this shift is monumental. Starting from July 7, the tech giant began transitioning key Microsoft 365 applications, such as Excel and Outlook, to its MAI models, effectively reducing its reliance on external AI services.
Moreover, MAI-Thinking-1’s 256K context window allows it to handle extensive inputs smoothly, making it ideal for complex documents like legal contracts. Independent benchmarks have shown that it consistently competes well against models from Anthropic, particularly in coding benchmarks.
The implications for the AI market are significant. As Microsoft reroutes prompts to its internal models, it could disrupt the revenue streams for OpenAI and Anthropic, altering the competitive landscape of AI deployment.
This material is informational and not financial advice.
