TechnologyTop Stories
GBO_Microsoft AI Models

Microsoft unveils MAI AI models to challenge OpenAI and Anthropic with cheaper coding and cybersecurity tools

Microsoft launches MAI Code 2, MAI Cyber 2 Flash and other AI models at Ignite, targeting OpenAI and Anthropic with lower-cost AI capabilities

Tech giant Microsoft on Tuesday announced a suite of AI models with powerful coding and cybersecurity capabilities in its bid to challenge rivals like Anthropic and OpenAI.

At the Microsoft Ignite conference in San Francisco, the company unveiled a computer programming model called MAI Code 2, which the Satya Nadella-led firm says almost matches the performance of rival AI models at a fraction of the cost.

The company has released new image-generation and transcription AI models as well.

“We need to parlay any technological advance into real outcomes and new points of leverage for people, organisations and society at large,” said Nadella in his opening keynote at Ignite.

“For us, that comes down to three things: keeping human ambition at the centre, ensuring you have ownership over your organisation’s intelligence and learning, and trust,” Nadella said.

Microsoft AI chief Mustafa Suleyman, in an interview with Reuters, said the company’s strategy of building technology specific to different domains was helping it keep costs low while ⁠matching rivals and, in some cases, topping model benchmarks.

“This isn’t just one model; it’s not a fleeting trend,” he said.

“This is a system, a hill-climbing machine that we have built, ⁠which co-optimises performance and price.”

Suleyman, taking the note of the recent cybersecurity incidents involving OpenAI agents, said the Hugging Face fiasco had led him to focus not ⁠just on performance but on critical safeguards as well.

“Preventing AI from taking harmful actions and ensuring it behaves as ⁠intended, a research area known as alignment, was his top priority,” he said.

Talking about the product line-up in detail, while MAI Cyber 2 Flash reportedly beats GPT-6 Astra’s quality at less than 2% of the cost per task on the CyberGymE2E benchmark, MAI Code 2, on the other hand, reaches 98% of the coding quality of GPT-6.1 Sol at about 40% of the price. And against Opus 5.5, it reaches 96% performance for just 5% of the price.

MAI Code 2 Flash, as per Microsoft, matches the performance of GPT-6 Luna at 65% of the price and Haiku 5.5 at 20% of the price.

“MAI Image 2.7 is number one on Artificial Analysis image editing, outperforming GPT Image 2.6 Sunburst (max) at just one-fifth of the price,” the tech giant claimed.

MAI Realtime 1 outperforms GPT-Realtime 2.1 and starts at just USD 0.01 per audio minute.

Dragon Copilot Transcribe has been built on MAI Transcribe 2, the world’s fastest and most accurate multilingual transcription model. It already powers 10 million patient encounters a month, with multilingual error rates halved.

Decision 1.1 is the latest decision model from Microsoft, 50x faster and 79x cheaper than GPT-6.1 Sol.

And last but not least, Maia 200, Microsoft’s second-generation AI accelerator, is now in production, running the company’s largest models, achieving 1.7x more tokens per watt.

Related posts

Nigerian logitech startup ShapShap raises funding from GreenTec Capital

GBO Correspondent

dHedge receives funds worth $1.15 mn from Framework Ventures, others

GBO Correspondent

Dubai implements expatriate saving scheme for govt staff

GBO Correspondent