# MAI Code 2

> Microsoft AI's coding model for everyday enterprise work, announced 17 November 2026 and available in GitHub Copilot.

- Maker: Microsoft
- Line: MAI Code (coding models)
- Canonical page: https://ai-tldr.dev/models/mai-code-2/
- Updated: 2026-11-17

## In short

MAI Code 2 is a coding model from Microsoft AI, announced on 17 November 2026 together with MAI Code 2 Flash. Microsoft aims it at common enterprise coding tasks: taking code to production, deep debugging, multi-file implementation, repo-wide reasoning and UI work.

## Facts

| Fact | Value | Source |
| --- | --- | --- |
| Licence | Proprietary · API only | [1] |
| Released | 17 November 2026 | [2] |

## Sources

1. [Microsoft AI — MAI Code 2 model page](https://microsoft.ai/models/mai-code-2/)
2. [Microsoft AI — Introducing MAI Code 2 and MAI Code 2 Flash](https://microsoft.ai/news/introducing-mai-code-2-and-mai-code-2-flash/)

## Overview

MAI Code 2 is a coding model from Microsoft AI, announced on 17 November 2026 together with MAI Code 2 Flash. Microsoft aims it at common enterprise coding tasks: taking code to production, deep debugging, multi-file implementation, repo-wide reasoning and UI work.

Microsoft tested it on DeepSWE 1.1, Terminal Bench 2.1, SWE-Atlas and FrontierCode v1.1, at max reasoning effort. Averaged over those four, it says MAI Code 2 reaches about 98% of GPT-6.1 Sol's performance at 40% of its cost, and 96% of Claude Opus 5.5's performance at 5% of its cost. In blind frontend evaluations with 258 annotators, reviewers preferred it over GPT-6 Sol 67% of the time. These are Microsoft's own results, and the cost figures reflect a 50% promotional price.

MAI Code 2 is available in GitHub Copilot. Microsoft's launch post does not state a context window, parameter count or per-token price, and does not say whether the weights are open.

## Strengths

- About 98% of GPT-6.1 Sol's averaged benchmark score at 40% of its cost, per Microsoft
- About 96% of Claude Opus 5.5's averaged score at 5% of its cost, per Microsoft
- Preferred over GPT-6 Sol 67% of the time in blind frontend evaluations with 258 annotators
- Available in GitHub Copilot at launch

## Best for

- Reach for it for multi-file implementation and repo-wide reasoning in GitHub Copilot
- Reach for it for deep debugging and taking code to production
- Reach for MAI Code 2 Flash instead for high-volume routine edits and fixes

## FAQ

### Where can I use MAI Code 2?

MAI Code 2 is available in GitHub Copilot. Microsoft also lists OpenRouter, Vercel, Hermes, Open Code and Kilo as coming soon, and says MAI Code 2 Flash is in limited-access preview on Microsoft Foundry.

### How much does MAI Code 2 cost?

Microsoft's launch post gives no per-token price. It states cost relative to rivals, about 40% of GPT-6.1 Sol and about 5% of Claude Opus 5.5, at a 50% promotional price.

### Are MAI Code 2's weights open?

Microsoft's launch post does not say. A separate offer, MAI Licensed Weights, lets customers customize frontier MAI models in their own Azure environment.

## Links

- [Microsoft AI — Introducing MAI Code 2 and MAI Code 2 Flash](https://microsoft.ai/news/introducing-mai-code-2-and-mai-code-2-flash/)
- [Microsoft AI — MAI Code 2 model page](https://microsoft.ai/models/mai-code-2/)
- [Microsoft AI — Ignite 2026 models roundup](https://microsoft.ai/news/microsoft-ignite-2026/)

Source: AI/TLDR LLM registry, https://ai-tldr.dev/models/mai-code-2/
