The six models
On November 17, 2026 Microsoft AI announced new MAI models at Ignite. Axios counted six, covering coding, cybersecurity, images and voice: MAI Code 2, MAI Code 2 Flash, MAI Cyber 2 Flash, MAI Image 2.7, MAI Realtime 1 and MAI Transcribe 2. Microsoft's own Ignite page lists more items, including Decision 1.1 and the Maia 200 chip. Everything below is the vendor's claim.
| Model | What Microsoft says | Vendor price | On AnyRouter |
|---|---|---|---|
| MAI Code 2 | 98% of GPT 6.1 Sol's coding performance at 40% of the cost, in GitHub Copilot | Not published | Coming soon |
| MAI Code 2 Flash | Comparable to GPT-6 Luna and Claude Haiku 5.5 at 65% and 20% of their cost | Not published | Coming soon |
| MAI Cyber 2 Flash | 92.3% on CyberGym-E2E against 91.5% for GPT-6 Astra, 1M-token context | $0.10 in, $0.50 out per 1M (offline estimate, not a customer price) | Coming soon |
| MAI Image 2.7 | First on Artificial Analysis image editing, by 4 Elo; public preview in Microsoft Foundry | About $35 per 1,000 images; Flash about $10 | Not listed |
| MAI Realtime 1 | Speech-to-speech, compatible with the OpenAI Realtime API; MAI Labs and a limited Foundry preview | From $0.01 per audio minute | Not listed |
| MAI Transcribe 2 | 60 languages, 3.4% average WER on FLEURS (top 25 languages); limited preview | $0.10 per hour of audio, limited time | Not listed |
Which ones AnyRouter lists
AnyRouter lists text and embedding models only, so the image, realtime voice and transcription models stay off the catalog. The three text models are listed as Coming soon, because Microsoft publishes no public API route or wire id for them yet. A request to one of them returns a model-not-yet-available response.
The two coding models share their own post with prices, benchmarks and plans: MAI Code 2 and MAI Code 2 Flash. MAI Cyber 2 Flash comes exclusively to Microsoft security products, including Codename MDASH, for eligible customers from mid-December, and Microsoft promises a GPT-6 Cyber comparison in the coming weeks.
One related model is callable today: Microsoft-Decision-1.1, released the same day, runs through your own Foundry or Vercel key.
How to read the numbers
Every benchmark and cost ratio above comes from Microsoft's announcements, run on its own setups. Microsoft says Code 2 and Code 2 Flash costs are at a 50% promotional price, and says its security results reflect specific tasks and configurations. Treat them as claims until a host serves the models and you can measure them on your own work.
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.