Business of GTM

Microsoft Prices MAI Code 2 at 40% of OpenAI's Sol on a Half-Price Promotion

Microsoft says Code 2 delivers 98% of the performance of OpenAI's GPT-6.1 Sol at 40% of the cost. FourWeekMBA estimates the cost at about 80% of Sol's if the promotional price doubles.

Black-and-white halftone portrait of Microsoft AI CEO Mustafa Suleyman speaking onstage with both hands raised, set against a red arc on a beige background.
Mustafa Suleyman, CEO of Microsoft AI, speaking at Microsoft Ignite 2026 in San Francisco in November 2026. Photo illustration by Business of GTM. Original photo courtesy of Microsoft.

TL;DR

  • Microsoft released seven of its own MAI models at Ignite on November 17, including two coding models inside GitHub Copilot.
  • The cost lead Microsoft is selling for MAI Code 2 rests on a half-price promotion.
  • The promotional Code 2 rates are scheduled to double on May 17, 2027, according to iTWire.

On November 17, Microsoft launched seven MAI models at Ignite for coding, cybersecurity, image generation and transcription, and pitched them on cost. The coding pair, MAI Code 2 and Code 2 Flash, reached GitHub Copilot the same day. Microsoft AI CEO Mustafa Suleyman framed the lineup around specialized models and their cost in comments to Reuters, as CXOToday reported.

The cost case for Code 2 depends on a promotional price that Microsoft has scheduled to raise. If that price doubles, FourWeekMBA estimates the saving against OpenAI's Sol shrinks to about a fifth of the bill.

The 40% figure assumes a half-price promotion

Microsoft says Code 2 delivers 98% of the performance of OpenAI's GPT-6.1 Sol. It puts Code 2's cost at 40% of Sol's. That comparison averages four coding benchmarks, runs every model at maximum reasoning effort and prices Code 2 at a 50% promotional discount, iTWire reports. The promotional rates are $0.50 per million input tokens and $2.50 per million output tokens.

FourWeekMBA estimates that if Microsoft's promotional price is simply doubled and token counts don't change, Code 2 would cost about 80% of what Sol does. Microsoft hasn't published a list price for Code 2. FourWeekMBA puts Code 2 Flash at about 130% of the cost of OpenAI's GPT-6 Luna and found no independent benchmark results for the new models.

Copilot switches the new models on by default

GitHub turns new models on for Copilot Business and Enterprise accounts automatically unless an admin switches them off. Code 2 is on the Pro, Pro+, Max, Business and Enterprise plans, and Flash adds Free and Student. Both are billed by usage, the same move Microsoft, HubSpot and Zoom made when they shifted AI pricing from the seat to the spend limit. GitHub plans to retire MAI-Code-1.1-Flash on December 18, according to iTWire.

Outside Copilot, Code 2 Flash is a limited preview in Foundry. Cyber 2 Flash is sold only inside Microsoft's security products, at list rates of $0.10 per million input tokens and $0.50 per million output tokens. It replaces the existing model for eligible Codename MDASH customers from mid-December. Microsoft dropped 92 of 905 cyber benchmark tasks that overlap its training data.

Analyst Patrick Moorhead wrote on X that cost was the keynote and the model is now a component. In Moorhead's view, Copilot's edge against ChatGPT is everyday use bundled into the seat, and the 98% figure is Microsoft's own number. That suggests the promotion's job is to start usage inside accounts Microsoft already bills.

The Code 2 discount ends on a date Microsoft already set

Copilot Business and Enterprise accounts that leave the setting on can run Code 2 at the promotional rate from launch day. Those rates are scheduled to double on May 17, 2027, according to iTWire.

Copilot's edge against ChatGPT is everyday use bundled into the seat, and the 98% figure is Microsoft's own number.

Patrick Moorhead, analyst

GitHub plans to retire MAI-Code-1.1-Flash on December 18, and Code 2's promotional rates of $0.50 and $2.50 per million tokens are scheduled to double on May 17, 2027.

By Business of GTM