↓ Skip to main content
9 – 17 UHR +49 8031 3508270 LUITPOLDSTR. 9, 83022 ROSENHEIM
DE / EN
LLM Microsoft USA

Microsoft MAI

Microsoft MAI: MAI Code 2, MAI-Image-2.7, MAI Realtime 1, MAI-Transcribe-2 and MAI-Thinking-1 with pricing, access routes, EU regions and GDPR notes.

License Proprietary (Microsoft)
GDPR Hosting Available
Context 1M (MAI Code 2, MAI Code 2 Flash, MAI Cyber 2 Flash), 256k (MAI-Thinking-1), 32k (MAI-Image) Tokens
Modality Text, Image, Audio → Text, Image, Audio

Versions

Overview of available model variants

ModelReleaseEUStrengthsWeaknessesStatus
MAI Code 2 Recommended
17 November 2026 (GitHub Copilot)
Per Microsoft 98% of GPT-6.1 Sol's performance at about 40% of the price and 96% of Claude Opus 5.5's performance at 5% of the price (average of DeepSWE 1.1, Terminal Bench 2.1, SWE-Atlas and FrontierCode v1.1) In a blind test with 258 raters, developers preferred its frontend results over GPT-6 Sol in 64% of cases, per Microsoft 1M-token context and configurable reasoning per the GitHub documentation Built for production-ready code, debugging, multi-file changes and repo-wide reasoning
Per Microsoft not tuned for fundamental science research or very long-horizon agentic work Only via GitHub Copilot, not in Copilot Free and Student Promotional price is time-limited; benchmarks come from Microsoft and ran on Microsoft's internal platform MSBench
Current
MAI Code 2 Flash
17 November 2026
Per Microsoft performance comparable to GPT-6 Luna at 65% and to Claude Haiku 5.5 at 20% of the price 1M-token context, up to 128K output tokens, text and image input (Foundry model catalog) Part of automatic model selection in GitHub Copilot Per Microsoft intended as a base model for adaptation to your own codebases; customer examples from AT&T and Woven by Toyota
In Foundry only with approval (limited access preview) Intended for routine work: explanations, targeted edits, simple fixes No open weights
Current
MAI-Image-2.7 / MAI-Image-2.7-Flash Recommended
17 November 2026 in Foundry (public preview)
Per Microsoft #1 on Artificial Analysis AA-Image-Editing v2.1 with 1,215 Elo (5,478 votes, as of 16 November 2026), 4 Elo ahead of GPT Image 2.6 Sunburst (max) at one fifth of the price For text-to-image Microsoft cites rank 4 Text-to-image and image editing with up to five reference images, targeted changes to colour, material or position without touching the rest of the image Transparent backgrounds for product images, logos and icons
Public preview without an SLA Per Microsoft Learn Global Standard only, no documented EU data zone Photorealistic edits involving minors blocked by default (access on request)
Preview
MAI Realtime 1
17 November 2026 (limited access preview)
Microsoft AI's first speech-to-speech model for real-time conversations with interruptions and quick acknowledgements Compatible with the OpenAI Realtime API, connects to enterprise systems and tools Per Microsoft better than GPT Realtime 2.1 on quality, speed and price (Big Bench Audio, τ-Voice, Dynamics 365 evaluations) Planned for Dynamics 365 Contact Center and Dragon Copilot
Approval required; no region or language details Microsoft publishes no concrete benchmark values
Preview
MAI Cyber 2 Flash
17 November 2026 (rollout from mid-December 2026)
1M-token context, four times that of MAI Cyber 1.1 Flash Per Microsoft 92.3% on CyberGym-E2E (GPT-6 Astra: 91.5%) with half the reasoning budget and under 2% of the cost per task Per Microsoft 40.4% on WebExploitBench (GPT-5.6 Cyber: 38.4%) and 86% fewer output tokens than MAI Cyber 1.1 Flash
Usable only inside Microsoft security products, not as a standalone API All performance figures come from Microsoft
Preview
MAI-Thinking-1 Recommended
2 June 2026 (public preview)
Microsoft's first fully in-house reasoning model, trained on commercially licensed data per Microsoft Sparse mixture-of-experts with 35B active parameters, 256k context, up to 64k output Per Microsoft 97.0% AIME 2025, 94.5% AIME 2026 and 52.8% SWE-Bench Pro OpenAI-compatible chat completions API (/mai/v1), function calling, encrypted chain of thought
Public preview without an SLA Global Standard only, no EU data zone; text input only
Preview
MAI-Transcribe-2 Recommended
3 September 2026 (public preview)
60 languages including German, multilingual mode with automatic language identification Per Microsoft an average word error rate of 5.2% on FLEURS across 60 languages Speaker diarization, word timestamps, keyword biasing, code switching, verbatim or clean style Foundation of Dragon Copilot Transcribe for doctor-patient documentation
Public preview without an SLA, not recommended for production workloads per Microsoft No Germany West Central for the fast transcription API Promotional price limited until 31 December 2026
Preview
MAI-Transcribe-2-Streaming
October 2026 (public preview)
Real-time transcription with continuous intermediate and final results Multilingual mode including German For call centers, voice assistants, live captions and real-time notes
Public preview without an SLA; the Microsoft Foundry retirement schedule lists 4 December 2026 for this preview version In the EU only Sweden Central; via Voice Live with a global processing scope there
Preview Shutdown: 4 December 2026 (preview, per the Microsoft Foundry retirement schedule)
MAI-Voice-2.1 / MAI-Voice-2.1-Flash
October 2026 (public preview)
23 languages; German voices de-DE-Klaus, de-DE-Mia, de-DE-Grant and de-DE-Harper Emotions and styles via SSML (mstts:express-as) Flash for low-latency voice agents and IVR, 2.1 for audiobooks, podcasts and educational content Instant voice cloning from 5 to 60 seconds of reference audio after limited-access approval
Public preview without an SLA Per the MAI-Voice page served globally with routing between the serving regions; no Germany West Central
Preview
MAI-Image-2.6 / MAI-Image-2.6-Flash
31 July 2026 (GA)
Text-to-image and image editing, up to 2,359,296 pixels (equivalent to 1536 × 1536) Automatic aspect ratio selection and web grounding via Bing as options
Superseded by MAI-Image-2.7; prompts in English only per the model catalog Global Standard only; GA per the retirement schedule, retiring on 1 May 2027
Current
MAI-Transcribe-1.5
2 June 2026
43 languages including German
Superseded by MAI-Transcribe-2: fewer languages
Current
MAI Cyber 1.1 Flash
27 October 2026 (MDASH)
Per Microsoft comparable results to MAI Cyber 1 Flash at 18% lower cost per scan
Replaced by MAI Cyber 2 Flash for eligible customers from mid-December 2026
Current
Older, deprecated and retired versions (5)

Legacy and deprecated: superseded by newer models; for many versions the provider has already announced a shutdown. Retired: no longer available. Dates and successors are in the model lifecycle section.

ModelReleaseEUStrengthsWeaknessesStatus
MAI-Code-1.1-Flash
11 August 2026 (GitHub Copilot)
Image input, 256K context
Per the GitHub changelog deprecated across all Copilot experiences on 18 December 2026; the replacement is the latest MAI Code Flash model
Deprecated Shutdown: 18 December 2026 (GitHub Copilot)
MAI-Image-2.5 / 2.5-Flash / 2.5-Pro
2 June 2026 (Pro: 19 June 2026), preview
Pro variant for photorealistic output and object consistency
Preview; retired on 24 November 2026 per the Microsoft Foundry retirement schedule (as of 18 November 2026)
Deprecated Shutdown: 24 November 2026 (per the retirement schedule)
MAI-Transcribe-1
January 2026 (preview)
First Transcribe model of the MAI family
Per Microsoft Learn deprecated since 20 August 2026, retired on 15 September 2026 per the retirement schedule
Retired Shutdown: 15 September 2026 (Microsoft Foundry)
MAI-Code-1-Flash
2 June 2026
Coding model from Build 2026, optimised for GitHub Copilot
Per the GitHub changelog retired across all Copilot experiences on 10 September 2026
Retired Shutdown: 10 September 2026 (GitHub Copilot)
MAI-Image-2 / MAI-Image-2e
February 2026 (2), April 2026 (2e)
First image models of the MAI family in Microsoft Foundry
MAI-Image-2e retired on 28 August 2026 per the Microsoft Foundry retirement schedule; the schedule no longer lists MAI-Image-2
Retired Shutdown: 28 August 2026 (MAI-Image-2e, Microsoft Foundry)

Use Cases

Typical applications for this model

Coding assistants in GitHub Copilot
Image generation and editing
Transcription and captions
Speech output and voice agents
Reasoning workloads in Microsoft Foundry
Vulnerability analysis in Microsoft security products

Technical Details

API, features and capabilities

API & Availability
Availability Microsoft Foundry (MAI-Thinking-1, MAI-Image, MAI-Code-2-Flash and MAI Realtime 1 as limited access previews), Azure Speech (MAI-Transcribe, MAI-Voice), GitHub Copilot (MAI Code 2, MAI Code 2 Flash), Microsoft security products (MAI Cyber)
Features & Capabilities
Tool Use Function Calling Vision Reasoning Mode Realtime API
Training & Knowledge
Knowledge Cutoff not documented
Fine-Tuning Available (Reinforcement fine-tuning for MAI-Code-1.1-Flash (preview, Global), MAI Licensed Weights and Frontier Tuning (announced))
Language Support
Best Quality English, German
Supported MAI-Transcribe-2: 60 languages, MAI-Voice-2.1: 23 languages; no language list published for MAI Code 2, MAI Realtime 1 and MAI-Image-2.7
German is documented for Transcribe and Voice; plan your own tests with German content for the other models

Hosting & Compliance

GDPR-compliant hosting options and licensing

GDPR-Compliant Hosting Options
Azure Speech (fast transcription)
MAI-Transcribe-2: North Europe, West Europe
Per Microsoft Learn, Azure Speech stores and processes speech data in the region of the resource; preview without an SLA
Azure Speech (Voice Live)
MAI-Transcribe-2 in EU resource regions such as Germany West Central, France Central, Sweden Central and West Europe
Processing scope Data Zone per Microsoft; LLM processing in Voice Live has its own processing scope
License & Hosting
License Proprietary (Microsoft)
Security Filters Microsoft Foundry guardrails and abuse monitoring
Enterprise Support Yes
Cloud Only

Benchmarks

Performance comparison with standardized tests

MAI Code 2: share of GPT-6.1 Sol's performance
98%
CyberGym-E2E
92.3%
WebExploitBench
40.4%
AIME 2025
97.0%
SWE-Bench Pro
52.8%

MAI (Microsoft AI) is Microsoft’s family of fully in-house models for software development, reasoning, images, speech and cybersecurity. At Ignite 2026 on 17 November 2026, Microsoft AI announced seven new models, including MAI Code 2, MAI Code 2 Flash, MAI-Image-2.7, MAI Realtime 1 and MAI Cyber 2 Flash. Unlike the open Phi models, all MAI models are proprietary and run through Microsoft services: Microsoft Foundry, Azure Speech, GitHub Copilot or Microsoft security products.

This page is based on the Microsoft AI Ignite overview, the Ignite 2026 Book of News, the model overview in Microsoft Learn, the GitHub price list and the Azure Retail Prices API. As an AI consultancy from Rosenheim we classify the models for enterprises in the DACH region by access route and EU processing. This overview reflects the state as of 19 November 2026.

Current models at a glance

ModelTypeStatusAccess routeEU per Microsoft
MAI Code 2CodingGA since 17 November 2026GitHub Copilotno EU processing commitment documented
MAI Code 2 FlashCodingGA in Copilot, limited access preview in FoundryGitHub Copilot, Microsoft FoundryFoundry: only Global prices so far
MAI-Image-2.7 / -2.7-FlashImage generation and editingPublic previewMicrosoft Foundry, fal, OpenRouter, VercelGlobal Standard
MAI Realtime 1Speech-to-speechLimited access previewMicrosoft Foundry, MAI Labsno details
MAI Cyber 2 FlashCybersecurityRollout from mid-December 2026Codename MDASHno details
MAI-Thinking-1Reasoning LLMPublic previewMicrosoft FoundryGlobal Standard, all EU regions
MAI-Transcribe-2Speech-to-textPublic previewAzure SpeechNorth Europe, West Europe
MAI-Transcribe-2-StreamingReal-time speech-to-textPublic previewAzure SpeechSweden Central
MAI-Voice-2.1 / -2.1-FlashText-to-speechPublic previewAzure SpeechFrance Central, North Europe, Sweden Central, West Europe
MAI-Image-2.6 / -2.6-FlashImage generationGAMicrosoft FoundryGlobal Standard, Sweden Central, West Europe

We cover the decision model Microsoft-Decision-1.1, also part of the Ignite announcements, on its own page Microsoft-Decision-1.1.

MAI Code 2 and MAI Code 2 Flash

MAI Code 2 targets everyday enterprise development work: production-ready code, debugging, multi-file changes, repo-wide reasoning and frontend work. Per Microsoft the model reaches 98% of GPT-6.1 Sol’s performance at about 40% of the price and 96% of Claude Opus 5.5’s performance at 5% of the price, averaged over DeepSWE 1.1, Terminal Bench 2.1, SWE-Atlas and FrontierCode v1.1. In a blind test with 258 raters, developers preferred its frontend results over GPT-6 Sol in 64% of cases. Per Microsoft all models ran at maximum reasoning effort; Microsoft measured DeepSWE, Terminal Bench and SWE-Atlas on its internal platform MSBench via the Copilot CLI, while FrontierCode comes from Artificial Analysis. Microsoft calculates costs at the promotional price. Per Microsoft the model is not tuned for fundamental science research or very long-horizon agentic work.

MAI Code 2 Flash is meant for high-volume routine tasks: code explanations, targeted edits, simple fixes and fast iterations. Microsoft cites performance comparable to GPT-6 Luna at 65% and to Claude Haiku 5.5 at 20% of the price. The Foundry model catalog lists the model as mai-code-2-flash (version 2026-11-17) with a 1M-token context, up to 128K output tokens, and text and image input. Microsoft has also derived MAI Work Flash from Code 2 Flash, which Microsoft uses in Microsoft 365 and Fabric IQ products.

Per the GitHub changelog and the GitHub pricing list, both models have been generally available in GitHub Copilot since 17 November 2026: MAI Code 2 on Pro, Pro+, Max, Business and Enterprise, MAI Code 2 Flash also on Free and Student. They are available among others in VS Code, Visual Studio, JetBrains, Xcode, Eclipse, the Copilot CLI, the Copilot cloud agent and on github.com. In Business and Enterprise, new models are enabled by default unless administrators have turned off the global default. In Microsoft Foundry only MAI Code 2 Flash is available, as a limited access preview.

MAI-Image-2.7 and MAI-Image-2.7-Flash

MAI-Image-2.7 and MAI-Image-2.7-Flash generate and edit images. Microsoft first showed 2.7 on 27 October 2026 as an image editing model outside Foundry; since Ignite both models are available as a public preview in Microsoft Foundry (model version 2026-11-02), and also via fal, Genspark, OpenRouter and Vercel. Per Microsoft, MAI Image is now the default model family of Bing Image Creator.

Per Microsoft, MAI-Image-2.7 ranks #1 on the Artificial Analysis image editing ranking AA-Image-Editing v2.1 with 1,215 Elo from 5,478 votes (as of 16 November 2026), 4 Elo ahead of GPT Image 2.6 Sunburst (max) at one fifth of the price. For text-to-image Microsoft cites rank 4. The models process a prompt of up to 32,000 tokens and up to five reference images, change the colour, material or position of individual objects in a targeted way and generate images with transparent backgrounds. Per Microsoft Learn, photorealistic edits involving minors are blocked by default.

MAI Realtime 1

MAI Realtime 1 is Microsoft AI’s first speech-to-speech model. It is designed to hold natural conversations with interruptions and quick acknowledgements, hand harder tasks to a stronger reasoning stage, understand requests despite background noise and connect to enterprise systems and tools. The model is compatible with the OpenAI Realtime API, which is meant to ease the migration of existing voice applications. Microsoft states that MAI Realtime 1 beats GPT Realtime 2.1 on quality, speed and price, but publishes no concrete values.

Per Microsoft the price starts at USD 0.01 per audio minute; reasoning costs USD 1 per 1M input tokens and USD 5 per 1M output tokens. Enterprises can register for the limited access preview in Microsoft Foundry and pay nothing until the public preview. A demo runs in MAI Labs, which per its terms of use is intended for personal, non-commercial purposes only. Microsoft plans to use it in Dynamics 365 Contact Center and Dragon Copilot.

MAI Cyber 2 Flash

MAI Cyber 2 Flash finds vulnerabilities in code, proves them and supports the investigation and detection of attacks. Microsoft offers the model exclusively in Microsoft security products, including Codename MDASH, for verified defenders. For eligible customers it replaces the existing model in the MDASH MAI profile from mid-December 2026; the Book of News refers to a preview.

Per Microsoft, MAI Cyber 2 Flash reaches 92.3% on CyberGym-E2E versus 91.5% for GPT-6 Astra, with half the reasoning budget and under 2% of the cost per task; Microsoft excluded 92 of the 905 tasks for all models because of overlap with training data. On WebExploitBench Microsoft cites 40.4% versus 38.4% for GPT-5.6 Cyber. The context grows fourfold over MAI Cyber 1.1 Flash to 1M tokens, and the model produces 86% fewer output tokens. The list price is USD 0.10 per 1M input tokens and USD 0.50 per 1M output tokens.

MAI Licensed Weights

MAI Licensed Weights is meant to let enterprises customize an MAI model with their own data and run it in their own Azure environment, with control over deployment, access and updates. Per Microsoft, data and adaptations remain with the customer; Microsoft uses them for training only with express agreement. As examples Microsoft cites AT&T (customized versions of MAI Code 2 and Code 2 Flash for 5G network analytics), Woven by Toyota (ready-to-merge fixes in 44% of the cases where other models needed rework) and Alaska Airlines (classification accuracy from 81.2% to 91.1%). The offering is announced and comes through Frontier Company engagements; Microsoft has not yet published dates, prices or licence terms. These are not open weights: the models still run in Azure.

MAI-Thinking-1

MAI-Thinking-1 (public preview since 2 June 2026) is the family’s reasoning model: sparse mixture-of-experts with 35B active parameters, a 256k context and up to 64k output tokens, trained on commercially licensed data per Microsoft. Microsoft cites 97.0% on AIME 2025, 94.5% on AIME 2026 and 52.8% on SWE-Bench Pro. In Microsoft Foundry the model runs via an OpenAI-compatible chat completions API (/mai/v1) with function calling and encrypted chain of thought, exclusively as Global Standard. The Azure price list shows USD 2.00 per 1M input tokens, USD 0.20 for cached input and USD 8.00 for output.

MAI-Transcribe-2 and MAI-Transcribe-2-Streaming

MAI-Transcribe-2 (public preview since 3 September 2026) transcribes 60 languages including German with speaker diarization, word timestamps, keyword biasing, automatic language identification and code switching; the style is selectable (verbatim with filler words or clean). Per Microsoft the average word error rate on FLEURS is 5.2%. The model is called via the Azure Speech fast transcription API (enhancedMode.model = "MAI-Transcribe-2") or as input transcription in Voice Live. The price is USD 0.10 per audio hour as a promotional offer until 31 December 2026. Dragon Copilot Transcribe is built on MAI-Transcribe-2; per Microsoft it transcribes all doctor-patient conversations in Dragon Copilot, about 10 million encounters a month.

MAI-Transcribe-2-Streaming is the real-time variant for call centers, voice assistants, live captions and real-time notes. It connects via the Azure Speech SDK or the Realtime API (OpenAI-Realtime-compatible WebSocket protocol). Per Microsoft Learn, Azure serves the model from Sweden Central, Central US, East US 2 and Southeast Asia.

MAI-Voice-2.1 and MAI-Voice-2.1-Flash

MAI-Voice-2.1 and MAI-Voice-2.1-Flash are the family’s text-to-speech models (public preview) with 23 languages, emotion and style control via SSML and instant voice cloning from 5 to 60 seconds of reference audio, which is available only after limited-access approval and with consent. The Flash variant targets low-latency voice agents, IVR and call centers, MAI-Voice-2.1 long-form content such as audiobooks, podcasts and educational content. For German, Microsoft Learn lists the voices de-DE-Klaus, de-DE-Mia, de-DE-Grant and de-DE-Harper.

Pricing

ModelInputCached inputOutputAccess and note
MAI Code 20.500.0252.50GitHub Copilot, promotional price until 17 May 2027, then 1.00 / 0.05 / 5.00
MAI Code 2 Flash0.050.0050.25GitHub Copilot and Foundry (Global), promotional price until 17 May 2027, then 0.10 / 0.01 / 0.50
MAI-Code-1.1-Flash0.200.021.20GitHub Copilot until 18 December 2026
MAI-Thinking-12.000.208.00Microsoft Foundry, Global Standard
MAI Cyber 2 Flash0.100.010.50List price per Microsoft AI, only in Microsoft security products

All prices in USD per 1M tokens. The image and speech models use their own units:

ModelPriceSource
MAI-Image-2.7about USD 35 per 1,000 images; per 1M tokens: text input 5.00, image input 8.00, image output 34.00 USDMicrosoft AI, Azure price list (Global)
MAI-Image-2.7-Flashabout USD 10 per 1,000 images; per 1M tokens: 1.75 / 2.50 / 10.00 USDMicrosoft AI, Azure price list (Global)
MAI-Image-2.6 / -2.6-Flashper 1M tokens: 5.00 / 8.00 / 38.00 USD and 1.75 / 2.50 / 19.00 USDAzure price list (Global)
MAI Realtime 1from USD 0.01 per audio minute, reasoning 1 / 5 USD per 1M tokensMicrosoft AI
MAI-Transcribe-2USD 0.10 per audio hour until 31 December 2026Azure Speech pricing page, Microsoft AI
MAI-Voice-2.1per 1M characters per the Azure Speech pricing pageAzure Speech pricing page

Prices as of 19 November 2026. The Azure price list also shows Data Zone prices for EU regions for MAI-Image-2.7 (for example USD 40.80 per 1M image output tokens instead of USD 34.00) and an hourly price of USD 0.54 for MAI-Transcribe-2. Because Microsoft Learn does not yet document the Data Zone deployment type for MAI-Image and Microsoft does not explain whether USD 0.54 is the price after the promotion ends, we check such values directly in the Foundry portal in projects.

EU availability and GDPR

For the GDPR assessment, access route and deployment type matter, not the model family:

Access routeModelsProcessing per Microsoft
Microsoft Foundry, Global StandardMAI-Thinking-1, MAI-Image-2.5 to 2.7Inference possible in any Azure region; stored data stays in the chosen geography
Azure Speech, fast transcriptionMAI-Transcribe-2, MAI-Transcribe-1.5Resource in North Europe or West Europe; per Microsoft Learn, Azure Speech processes in the resource region
Azure Speech, Voice LiveMAI-Transcribe-2, MAI-Transcribe-2-StreamingMAI-Transcribe-2 in EU regions with processing scope Data Zone, streaming in Sweden Central with processing scope Global
Azure Speech, text-to-speechMAI-Voice-2.1, -2.1-FlashServed from France Central, North Europe, Sweden Central and West Europe, among others, with routing between these regions
GitHub CopilotMAI Code 2, MAI Code 2 Flashno EU processing commitment documented for the models
Microsoft security products, MAI LabsMAI Cyber 2 Flash, MAI Realtime 1 (demo)no region details; MAI Labs for personal use only

For Foundry models Microsoft states: prompts and responses are not used to train the base models and are not available to other customers or model providers. Even with Global deployments, stored data, including the abuse monitoring data store, stays in the customer-designated geography; for deployments in the EEA, per Microsoft only employees located in the EEA review flagged content. New since 17 November 2026: for models made available on or after 1 November 2026, Microsoft intends to stop offering the Modified Abuse Monitoring program from a date not yet announced, with at least 30 days’ notice. Based on the version dates this affects MAI-Image-2.7 and MAI-Code-2-Flash, among others. In addition, per Microsoft preview models may follow different privacy practices.

Our assessment: for personal data, the speech models with a resource in an EU region are currently the most suitable option, for transcription for example MAI-Transcribe-2 in West Europe or North Europe. We use Global Standard models such as MAI-Thinking-1 and MAI-Image for content without personal data, or assess them case by case with a data protection impact assessment and a data processing agreement.

Model lifecycle

  • MAI-Code-1.1-Flash is deprecated across all GitHub Copilot experiences on 18 December 2026 per the GitHub changelog; the replacement is the latest MAI Code Flash model, i.e. MAI Code 2 Flash.
  • MAI-Code-1-Flash has been retired in GitHub Copilot since 10 September 2026.
  • MAI-Image-2.5, -2.5-Flash and -2.5-Pro retire on 24 November 2026 per the Microsoft Foundry retirement schedule (as of 18 November 2026). Successors are MAI-Image-2.6 and MAI-Image-2.7.
  • MAI-Image-2e has been retired since 28 August 2026 per the retirement schedule; MAI-Image-2 is no longer listed there.
  • MAI-Transcribe-2-Streaming is listed as a preview with a retirement date of 4 December 2026.
  • MAI-Transcribe-1 has been retired since 15 September 2026; successors are MAI-Transcribe-1.5 and MAI-Transcribe-2.
  • MAI-Voice-2 and -2-Flash are no longer listed in the current Microsoft Learn documentation; successors are MAI-Voice-2.1 and -2.1-Flash.
  • MAI Cyber 1.1 Flash is replaced by MAI Cyber 2 Flash for eligible MDASH customers from mid-December 2026.
  • Apart from the coding models in GitHub Copilot, most MAI models run as previews without an SLA.

Our recommendation

For coding assistants, MAI Code 2 in GitHub Copilot is worth testing when cost per developer matters: the promotional prices until May 2027 are well below those of the large frontier models. MAI Code 2 Flash suits routine tasks and teams on Copilot Free or Student. If you have pinned MAI-Code-1.1-Flash, switch before 18 December 2026. For code with high confidentiality requirements, we review Copilot policies and data flows first.

For image generation and product images, MAI-Image-2.7 in Microsoft Foundry is worth evaluating, and the Flash variant at about USD 10 per 1,000 images for high volumes. For transcription in the Microsoft stack, MAI-Transcribe-2 with a Speech resource in West Europe or North Europe remains our first choice for evaluations; we plan it for production only once it is generally available. For voice agents we are watching MAI Realtime 1 until Microsoft names languages, regions and a public preview. We use MAI-Thinking-1 for reasoning tasks without personal data.

Via the AI Gateway and CompanyGPT, MAI models from Microsoft Foundry can be connected centrally alongside other models, with budgets and model switching in one place. innFactory GmbH is a Microsoft Solutions Partner; we support selection, evaluation, data protection review and integration.

Related Microsoft models

Open models for self-hosting and edge are covered on the Microsoft Phi page, the decision model for routing and classification on the Microsoft-Decision-1.1 page. We cover the OpenAI models in Microsoft Foundry under OpenAI GPT.

Get Microsoft MAI through innFactory

Microsoft MAI is available via Microsoft Foundry. Which variants run in which regions is described in the EU availability section. innFactory GmbH is a cloud reseller for AWS, Microsoft Azure, Google Cloud and STACKIT: you can source the cloud resources for it directly through us.

  • Billing by invoice or direct debit, no credit card required – all cloud providers on one consolidated invoice.
  • A personal contact at innFactory for setup, EU regions and model selection.
  • innFactory AI Gateway: budgets, SSO and cost control per team and project, as an infrastructure component in your own tenant.
  • CompanyGPT: the GDPR-compliant AI platform for all employees, with free choice of model.

Cost estimation for this model

For up-to-date token pricing, model variants and EU availability, see our sister project ai-prices.eu. It helps you compare and estimate the operational cost of leading AI models for your specific use case.

Compare prices on ai-prices.eu

ai-prices.eu is a project by innFactory AI Consulting GmbH and provides transparent cost estimates for leading AI models.

Frequently Asked Questions

Which MAI models did Microsoft announce at Ignite 2026?

On 17 November 2026 Microsoft AI announced seven new models: MAI Code 2 and MAI Code 2 Flash for software development, MAI Cyber 2 Flash for cybersecurity, MAI Image 2.7 with its Flash variant for image generation and editing, MAI Realtime 1 for real-time voice conversations, Dragon Copilot Transcribe built on MAI-Transcribe-2, and the decision model Microsoft-Decision-1.1. It also announced MAI Licensed Weights, which is meant to let enterprises customize MAI models and run them in their own Azure environment.

What do MAI Code 2 and MAI Code 2 Flash cost?

Per the GitHub price list, MAI Code 2 in GitHub Copilot costs USD 0.50 per 1M input tokens, USD 0.025 for cached input and USD 2.50 for output. MAI Code 2 Flash costs USD 0.05, USD 0.005 and USD 0.25. Both are promotional prices until 17 May 2027; after that GitHub lists USD 1.00 / 0.05 / 5.00 for MAI Code 2 and USD 0.10 / 0.01 / 0.50 for MAI Code 2 Flash. MAI Code 2 Flash is also available as a limited access preview in Microsoft Foundry.

Is MAI-Image-2.7 available in Microsoft Foundry?

Yes. Since Ignite, MAI-Image-2.7 and MAI-Image-2.7-Flash run as a public preview in Microsoft Foundry (model version 2026-11-02, deployment type Global Standard), and also via fal, Genspark, OpenRouter and Vercel. Microsoft AI cites about USD 35 per 1,000 images for MAI-Image-2.7 and about USD 10 per 1,000 images for the Flash variant. Both models generate and edit images, accept up to five reference images and support transparent backgrounds.

Do the MAI models offer EU data residency?

It depends on the model. Microsoft Learn documents the Foundry models MAI-Thinking-1 and MAI-Image as Global Standard only: inference can run in any Azure region, while stored data stays in the chosen geography according to Microsoft. MAI-Transcribe-2 is available via Azure Speech in North Europe and West Europe, and for Voice Live Microsoft lists the processing scope Data Zone for EU regions. MAI Code 2 runs via GitHub Copilot without a documented EU processing commitment. The family is therefore not GDPR-compliant across the board; model, access route and region decide.

What is MAI Realtime 1?

MAI Realtime 1 is Microsoft AI's first speech-to-speech model for natural real-time voice conversations, compatible with the OpenAI Realtime API. Microsoft cites a starting price of USD 0.01 per audio minute, with reasoning billed at USD 1 per 1M input tokens and USD 5 per 1M output tokens. The model runs as a limited access preview in Microsoft Foundry with registration, and per Microsoft there are no charges until the public preview. Microsoft has not yet published languages or regions.

Who can use MAI Cyber 2 Flash?

Per Microsoft, MAI Cyber 2 Flash is available exclusively in Microsoft security products, including Codename MDASH, and targets verified defenders. For eligible customers it replaces the existing model in the MDASH MAI profile from mid-December 2026. Microsoft cites a 1M-token context, 92.3% on CyberGym-E2E and a list price of USD 0.10 per 1M input tokens and USD 0.50 per 1M output tokens.

Are MAI models open source? What are MAI Licensed Weights?

No, all MAI models are proprietary. MAI Licensed Weights is an announced offering that is meant to let enterprises customize an MAI model with their own data and run it in their own Azure environment, with control over access and updates. Per Microsoft, Microsoft uses the data and adaptations for training only with express agreement. The offering comes through Frontier Company engagements; Microsoft has not yet named dates or prices. If you need open weights, use Microsoft Phi.

What is MAI-Transcribe-2 and what does it cost?

MAI-Transcribe-2 is Microsoft AI's speech-to-text model, available as a public preview via Azure Speech in Microsoft Foundry since 3 September 2026. It transcribes 60 languages including German with speaker diarization, word timestamps, keyword biasing and code switching. The price is USD 0.10 per audio hour as a promotional offer until 31 December 2026. Dragon Copilot Transcribe, which Microsoft uses for doctor-patient documentation, is also built on MAI-Transcribe-2.

Does MAI-Voice-2.1 offer German voices?

Yes. Per Microsoft Learn, MAI-Voice-2.1 and MAI-Voice-2.1-Flash offer the German voices de-DE-Klaus, de-DE-Mia, de-DE-Grant and de-DE-Harper with emotions and styles controllable via SSML. Both models are public previews without an SLA and support 23 languages. Microsoft serves them from France Central, North Europe, Sweden Central and West Europe, among other regions; instant voice cloning is available only after limited-access approval.

Consultation for this model?

We help you select and integrate the right AI model for your use case.