Microsoft Unveils Seven New AI Models, Challenging Rivals
REDMOND, Wash. – Microsoft kicked off its annual Build developer conference Tuesday by announcing seven new artificial intelligence models, asserting that they surpass key benchmarks set by competitors like Anthropic and Google. The move signals Microsoft’s ambition to be recognized as a leading AI innovator, not just a major investor and infrastructure provider for OpenAI.
“Super excited to announce seven new world-class MAI models today,” Microsoft AI CEO Mustafa Suleyman posted on X. “They represent what we consider a new era in AI designed to keep you in control and on the frontier.”
At the forefront of this release is MAI-Thinking-1, a reasoning model that Microsoft is positioning as its flagship text foundation model.
According to Suleyman, independent evaluators found MAI-Thinking-1 to be preferred over Anthropic’s Claude Sonnet 4.6 in blind tests. He also stated that the model achieved a 97% score on AIME 2025, a benchmark designed to assess advanced problem-solving and reasoning capabilities.
Suleyman further noted that the model’s performance on the SWE Bench Pro places it “right alongside Opus 4.6 on one of the toughest coding benchmarks.”
The company also introduced MAI-Code-1-Flash, a streamlined coding model developed for integration with GitHub Copilot and Visual Studio Code. Additionally, MAI-Image-2.5 and its “Flash” variant are claimed to outperform Google’s Nano Banana Pro in image-editing tasks. Rounding out the new offerings are MAI Transcribe-1.5, a transcription model supporting 43 languages, and MAI-Voice-2, a speech-generation model capable of producing natural-sounding voices in 15 languages and adapting to a speaker’s voice from brief audio samples.
“This is an extraordinary time in technology. The compute used to train frontier models has increased by a factor of one trillion,” Suleyman stated in a separate blog post detailing the new models. “Now we expect another thousand-fold increase over the next three years, which in turn means more advanced capabilities, and the continued rollout of ever more effective AI.”
This announcement arrives amidst escalating competition among major AI developers. Last week, Anthropic revealed its latest flagship model, Opus 4.8, which the company reported as being faster and more capable on benchmark tests, along with a suite of new features. On Tuesday, Anthropic also announced an expansion of its Project Glasswing, granting 150 companies access to its new cybersecurity-focused Mythos model.
Meanwhile, at its Google I/O event in May, Google unveiled Gemini Omni, a multimodal AI model that integrates Gemini with the company’s Veo, Nano Banana, and Genie media-generation models. Google also introduced Gemini Spark, a cloud-based AI agent designed to manage tasks across applications and workflows on behalf of users.
Microsoft’s introduction of its own proprietary AI models suggests a strategic shift to build in-house capabilities, moving beyond its established reliance on OpenAI’s technology. The company indicated that its MAI models “delivered the highest win rate, outperforming GPT-5.5 on quality, while being 10x lower on cost.”
“Developers and businesses have been crying out for AI that delivers on their terms and under their say,” Suleyman wrote. “We see this as a major step towards delivering that.”
Leave a Reply