NewsStocksMicrosoft (MSFT) Unveils MAI Code 2, Taking Aim at Anthropic and OpenAI on Price

Microsoft (MSFT) Unveils MAI Code 2, Taking Aim at Anthropic and OpenAI on Price

Author: Blockonomi·

Key Takeaways

  • •Microsoft unveiled six MAI models spanning coding, cybersecurity, image generation, and voice at its Ignite conference in San Francisco.
  • •MAI Code 2 reportedly delivers 98% of GPT-6.1 Sol's performance at 40% of the cost and is now available through GitHub.
  • •The announcement landed on the same day OpenAI released GPT-6.2, highlighting Microsoft's simultaneous roles as OpenAI's partner and challenger.
  • •Mustafa Suleyman said building domain-specific models keeps costs low and that alignment has become his top priority after OpenAI agents hacked the Hugging Face platform.
  • •Meta's Muse Wonder, Muse Spark 1.3, and Muse Ember models are coming to Microsoft Foundry, though no specific launch dates were provided.
Microsoft (MSFT) Unveils MAI Code 2, Taking Aim at Anthropic and OpenAI on Price

Microsoft (MSFT) is trying to win the AI model race on price, not just power. At its Ignite conference in San Francisco on Tuesday, the company showed off a batch of new artificial intelligence models. The star of the lineup is MAI Code 2, a computer programming model that Microsoft says almost matches rival systems on performance at a fraction of the cost. The rivals in question are Anthropic and OpenAI, and Microsoft is racing to challenge both.

All told, Microsoft unveiled six MAI models at the event, spanning coding, cybersecurity, image generation and voice, according to its official Ignite 2026 announcement. The rollout landed on the same day that OpenAI, Microsoft's partner, released GPT-6.2, as summarized by Wall St Engine in a post on X:

MICROSOFT ANNOUNCES 6 NEW AI MODELS AT IGNITE $MSFT unveiled six MAI models across coding, cybersecurity, image generation and voice on the same day its partner OpenAI released GPT-6.2. 1) MAI Code 2: Delivers 98% of GPT-6.1 Sol's performance at 40% of the cost. Now in GitHub… pic.twitter.com/yn6qA2K17c

— Wall St Engine (@wallstengine) November 17, 2026 (post on X)

By that account, MAI Code 2 delivers 98% of the performance of GPT-6.1 Sol at 40% of the cost, and the model is now available in GitHub. That placement matters for reach: GitHub is Microsoft's own code-hosting platform, so the new model lands directly in front of the developers it targets. The same-day timing also put Microsoft's price-focused lineup and its partner's release in front of customers at once, a reminder that Microsoft is simultaneously OpenAI's partner and its challenger.

Coding was not the only launch. Microsoft also unveiled a model built just for cybersecurity. Most customers will not get it right away: the company plans to offer it to a limited group of customers. The goal is to deal with a surge in software vulnerabilities, and AI tools are increasingly bringing those flaws to light. How quickly Microsoft widens access beyond that group is one thing to watch as the rollout develops.

Microsoft also released new image-generation and transcription models. That makes four model launches in a single event.

The cost-cutting playbook

Mustafa Suleyman, Microsoft's AI chief, said the trick is going narrow. Building technology for specific domains keeps costs low, he told Reuters. In his telling, that approach lets Microsoft match rivals and, in some cases, top model benchmarks. Pricing has become a central competitive lever as companies move AI coding tools from experiments into everyday development work.

The company is not leaning on one giant model to do it. "This isn't just one model. It's not a flash in the pan," Suleyman said. He called the effort "a system, a hill-climbing machine" that co-optimizes performance and price. The idea is to run a toolbox rather than a single giant hammer: each model gets a job and a budget.

Safety gets a seat at the table

Suleyman said a recent cybersecurity reckoning changed his focus, pointing to OpenAI agents hacking the Hugging Face platform. He now looks at safeguards as well as performance, and alignment has become his top priority. Alignment is the research area aimed at stopping AI from taking harmful actions. It also means making sure the technology behaves as intended.

Meta models head to Microsoft Foundry

Ignite also brought news on outside models. CEO Satya Nadella said Meta's latest AI models are coming to Microsoft Foundry, Microsoft's platform for building AI applications and agents on Azure. "We're giving developers and enterprises even more choice and flexibility!" Nadella wrote on X.

In a separate post, Meta CEO Mark Zuckerberg said Muse Wonder, Muse Spark 1.3 and Muse Ember are "coming soon." No specific launch dates were given; their arrival would add to the third-party options on Foundry alongside Microsoft's own MAI lineup.

Source: Blockonomi