MAI Code 2 and MAI Code 2 Flash
Microsoft AI released two coding models, MAI Code 2 and Code 2 Flash, and says Code 2 gives 98% of GPT 6.1 Sol's performance at 40% of the cost.
MAI Code 2 targets everyday enterprise coding such as debugging, multi-file implementation, repo-wide reasoning and UI work. Code 2 Flash targets high-volume work such as code explanations, targeted edits and routine fixes. Microsoft says they are not yet tuned for harder problems like fundamental science research or very long horizon agentic work.
- Date
- Tuesday 17 November 2026
- Lab
- Microsoft AI
- Kind
- model
- Access
- closed API
- Price
- Promotional 50% discount until May 17, 2027; standard rates afterwards are $1 per million input tokens and $5 per million output tokens (per MobiDevices)
Figures
| Measure | Value | Measured by |
|---|---|---|
| Performance relative to GPT 6.1 Sol, averaged over DeepSWE 1.1, Terminal Bench 2.1, SWE-Atlas and FrontierCode v1.1 | 98% at 40% of the cost All models run at max reasoning effort on MSBench with GitHub Copilot CLI without subagents; Flash cost shown at its 50% promotional price | company |
| Performance relative to Claude Opus 5.5, same four-benchmark average | 96% at 5% of the cost Same setup as above | company |
Available in GitHub Copilot; Code 2 Flash is in limited-access preview in Microsoft Foundry; OpenRouter, Vercel, Hermes, Open Code and Kilo are promised soon. Flash has a 50% promotional price, but list prices are not stated on the page.
Sources
- microsoft.ai/news/introducing-mai-code-2-and-mai-code-2-flash/
- mobidevices.com/en/microsoft-mai-code-2
- thehindubusinessline.com/info-tech/microsoft-takes-on-openai-and-anthropic-with-new-coding
This record was checked against its sources on 17 November 2026. How we check