Microsoft has unveiled a family of seven in-house AI models at its annual Build developer conference in San Francisco, marking a strategic shift to reduce its dependence on AI companies it has heavily funded.
Mustafa Suleyman, Microsoft AI chief executive, said that after tuning for consulting firm McKinsey, the company outperformed OpenAI's GPT-5.5 on quality with ten times better cost efficiency.
"We believe the time has come for every company to move from consuming a frontier model to fully participating at the frontier," said CEO Satya Nadella.
The headline release is MAI-Thinking-1, Microsoft's first reasoning model, trained from scratch on clean data. A mid-sized model with 35 billion active parameters, it is designed for complex multi-step instructions and code generation.
Alongside it, the company launched MAI-Code-1-Flash, a coding model now rolling out across GitHub Copilot and Visual Studio Code. By running its own models on Azure, Microsoft can cut third-party fees and pass savings to developers.
In blind evaluations, MAI-Thinking-1 was preferred over Anthropic's Claude Sonnet 4.6 and matches Claude Opus 4.6 on coding benchmarks.
Quantum Leap
Microsoft also announced its Majorana 2 quantum chip is 1,000 times more reliable than its predecessor. Qubits now survive an average of 20 seconds, compared to milliseconds before.
"We will have a quantum machine in 2029 that can solve commercially viable problems," said Zulfi Alam, corporate vice president of Microsoft Quantum.
IPO Race Heats Up
Anthropic filed confidentially for an IPO on June 1 after raising $65 billion in Series H, pushing its valuation to $965 billion. OpenAI is also preparing its own confidential filing.
Microsoft has committed $13 billion to OpenAI and up to $5 billion in Anthropic.