Two models, one release
On November 17, 2026 Microsoft AI announced MAI Code 2 and MAI Code 2 Flash at Ignite. They are one family released on the same day, so they share this post. Each keeps its own model page and its own id.
| Model | Built for | Where it runs today | API price |
|---|---|---|---|
| MAI Code 2 | Deep debugging, multi-file implementation, repository-wide reasoning, UI work | GitHub Copilot | Not published |
| MAI Code 2 Flash | High-volume work: code explanations, targeted edits, routine fixes, fast iteration | GitHub Copilot, limited-access preview in Microsoft Foundry | Not published |
Microsoft states cost only relative to other models, and says the costs it shows are at a 50% promotional price. It publishes no per-token rate for either model, so AnyRouter shows "Not published" instead of a number.
What Microsoft claims
These are Microsoft's published results, not AnyRouter measurements. Microsoft ran all models at max reasoning effort, and ran its own results on its internal MSBench with the GitHub Copilot CLI and no subagents.
- MAI Code 2: across DeepSWE 1.1, Terminal Bench 2.1, SWE-Atlas and FrontierCode v1.1, 98% of GPT 6.1 Sol's performance at 40% of the cost, and 96% of Claude Opus 5.5's performance at 5% of the cost.
- MAI Code 2: in a blind frontend evaluation with 258 annotators, reviewers preferred its results over GPT-6 Sol 64% of the time.
- MAI Code 2 Flash: across three day-to-day coding benchmarks, performance comparable to GPT-6 Luna and Claude Haiku 5.5 at 65% and 20% of their cost.
- Limit Microsoft names: MAI Code 2 is not yet tuned for fundamental science research or very long-horizon agentic work.
Where you can use them today
Both models are in GitHub Copilot. MAI Code 2 is on Copilot Pro, Pro+, Max, Business and Enterprise. MAI Code 2 Flash is on those plans plus Free and Student.
Microsoft lists OpenRouter, Vercel, Hermes, Open Code and Kilo as coming soon. When a host serves a model with a public wire id and a price, AnyRouter can add the route.
Status on AnyRouter
Both rows are listed as Coming soon. Until a host serves them, a request to either id returns a model-not-yet-available response. This is what the call will look like once they are routable:
anyr claude --model "microsoft/mai-code-2"import { createAnyRouter } from "@anyr/ai-sdk-provider"
import { generateText } from "ai"
const anyrouter = createAnyRouter()
const { text } = await generateText({
model: anyrouter("microsoft/mai-code-2-flash"),
prompt: "Explain this function in two sentences.",
})Read the Microsoft announcement for the full claims and footnotes.
Related posts
- Run coding agents through one gatewayGuidesJuly 12, 2026 · Point Claude Code, Codex, OpenCode, Cursor, or any OpenAI-compatible tool at AnyRouter and get one key for every model, unified billing, per-request audit logs, and app attribution. Two environment variables, or one npx command with the anyr CLI launcher.
- Use AnyRouter with the Vercel AI SDKGuidesJuly 19, 2026 · @anyr/ai-sdk-provider wraps the AnyRouter gateway as a standard Vercel AI SDK provider. Call createAnyRouter(), pass a provider/model id, and generateText, streamText, and generateObject work unchanged, switching models is a one-string edit.
- What is AnyRouter? A plain-English guide to AI model routingGuidesAugust 5, 2026 · An AI model router lets you call many LLMs through one API. Here's what AnyRouter is, why it exists, and whether it's useful for your next project.
Route your first request in 2 minutes
Start free with your own keys, or top up and pay per token. Get $4/mo in credits and free models on Go — $2/mo, or free when you donate a provider key.