Thursday, 19 November 2026, story 2 of 7
Microsoft AI launches seven in-house models led by MAI Code 2
Microsoft says MAI Code 2 reaches 98% of GPT-6.1 Sol's performance at about 40% of the price, with Copilot pricing promotional until May 17, 2027.

What happened
Microsoft AI released seven models at Ignite, covering code, security, images, speech and decision routing. Microsoft says MAI Code 2 reaches 98% of GPT-6.1 Sol's coding performance at about 40% of the price, averaged over four benchmarks. This adds to the Opus 5.5 comparison reported yesterday.
In GitHub Copilot, MAI Code 2 costs $0.50 per million input tokens and $2.50 per million output tokens through May 17, 2027, then $1 and $5, The Decoder reports. Microsoft says the model is not yet tuned for harder work such as very long-horizon agentic tasks.
What it means
The cost ratios are Microsoft's own, from its benchmark mix. The Decoder reports that three of the four tests ran on Microsoft's internal MSBench platform, using the Copilot CLI without sub-agents, and that Microsoft applied a 50% promotional discount to its own model. If that discount is built into the 40% and 5% figures, the gap will likely narrow once the $1 and $5 rates start after May 17, 2027. The sources give no ratios at those rates.
The stated scope matters too. A model aimed at everyday enterprise coding fits routine edits and fixes, not long-running agents. Teams on Copilot can check the claim on their own repositories. Independent results, and a published list-price comparison, are what to watch.
What to do: Test MAI Code 2 on your own repositories in Copilot, and price the work at the post-promotion list rates of $1 and $5, not today's.
Sources
Earlier on this: Microsoft's MAI Code 2 models arrive in GitHub Copilot (18 November 2026).
Filed under Models. Published .
"What it means" is analysis, not reporting. Found an error? Write to daily@vibitely.com.