# MAI-Cyber-1-Flash [Canonical model page](https://headlessthrones.com/models/microsoft-mai-cyber-1-flash) · [Model library](https://headlessthrones.com/models) Public catalog reference from Headless Thrones. Exact catalog ID: `microsoft-mai-cyber-1-flash`. Catalog source-check date: 2026-10-23. This brief contains the public catalog answers and any catalog guide, not private drafts, historical editorial revisions or independently measured rankings. The canonical HTML page may also display separately published editorial. Current catalog answers ## What to know about MAI-Cyber-1-Flash AI-compiled from linked model documentation. This is a catalog reference, not an independent performance review. Catalog checked October 23, 2026. [Exact model](https://headlessthrones.com/models/microsoft-mai-cyber-1-flash#model-answer-identity) · [Access](https://headlessthrones.com/models/microsoft-mai-cyber-1-flash#model-answer-access) · [Token limits](https://headlessthrones.com/models/microsoft-mai-cyber-1-flash#model-answer-context) · [Inputs & outputs](https://headlessthrones.com/models/microsoft-mai-cyber-1-flash#model-answer-modalities) · [Model details](https://headlessthrones.com/models/microsoft-mai-cyber-1-flash#model-answer-research) ### Which MAI-Cyber-1-Flash does this page describe? MAI-Cyber-1-Flash is a model record from **Microsoft** in the MAI family. [Source](https://microsoft.ai/models/mai-cyber-1-flash/) The recorded model identifier or checkpoint is `MAI-Cyber-1-Flash`. Use that exact identifier when checking documentation or a provider listing. [Source](https://microsoft.ai/pdf/MAI-Cyber-1-Flash-Model-Card.pdf) ### How can you access MAI-Cyber-1-Flash? The catalog documents **MDASH private preview for approved defenders** access for MAI-Cyber-1-Flash. [Source](https://microsoft.ai/pdf/MAI-Cyber-1-Flash-Model-Card.pdf) As of the catalog check, this version was listed without confirmed general availability. [Source](https://microsoft.ai/models/mai-cyber-1-flash/) ### What are MAI-Cyber-1-Flash's context and output limits? The publisher reports the context window as **256k tokens**. [Source](https://microsoft.ai/pdf/MAI-Cyber-1-Flash-Model-Card.pdf) We retain the publisher's notation and qualifiers; no exact token count or hosted serving allowance is inferred from it. A separate maximum-output value is not established in this record. The context or input figure should not be used as an output allowance. Context, input and output limits describe different constraints. Do not add the figures together to infer a larger window. Check the endpoint's accounting for conversation history, system instructions, tools, media and generated tokens; a documented maximum is not a guarantee of answer quality or a measured usable capacity. ### What can MAI-Cyber-1-Flash receive and generate? Documented inputs: **text**. [Source](https://microsoft.ai/models/mai-cyber-1-flash/) Documented outputs: **text**. [Source](https://microsoft.ai/models/mai-cyber-1-flash/) Input support and output generation are separate capabilities. Reading an image does not establish image generation; a text output listing does not establish speech output. File sizes, formats and media limits depend on the documented implementation. ### What distinguishes MAI-Cyber-1-Flash? These details come from the linked publisher material for this model. Reported capabilities and training descriptions are publisher statements, not independent test results. **Restricted access**: Microsoft makes this defensive cybersecurity model available only to verified defenders through MDASH; a public standalone model API is not established by this page. [Source](https://microsoft.ai/models/mai-cyber-1-flash/) **Deployment form**: MAI-Cyber-1-Flash is integrated into MDASH, a multi-agent system for identifying, validating and remediating software vulnerabilities. [Source](https://microsoft.ai/models/mai-cyber-1-flash/) **Specialized architecture**: Microsoft describes this cybersecurity fine-tune of MAI-Code-1-Flash as a mixture-of-experts model with 137B total and 5B active parameters. Active parameters are not a memory requirement. [Source](https://microsoft.ai/pdf/MAI-Cyber-1-Flash-Model-Card.pdf) **Deployment boundary**: The model card restricts integration to codename MDASH. This is an approved private-preview security workflow, not a standalone general-purpose API offer. [Source](https://microsoft.ai/pdf/MAI-Cyber-1-Flash-Model-Card.pdf) **Language evaluation limits**: The model card warns that its training and testing were primarily on English data and benchmarks, so performance for non-English customers may be lower. Language support is not evidence of equal evaluated quality. [Source](https://microsoft.ai/pdf/MAI-Cyber-1-Flash-Model-Card.pdf) **Catalog qualifications:** Documented specialist language/code model used within MDASH; general access was not verified. **Evidence still missing here:** knowledge and training-data cutoffs. Missing documentation is not evidence that a feature is unavailable. Headless Thrones has not independently benchmarked this model for this reference. Evaluate it on representative work before choosing it; these facts do not rank its accuracy, reliability, speed or suitability against another model. Keep the exact version in view ## Other MAI models from Microsoft. These are separate records in the same publisher family. A shared family name does not establish equal capabilities or interchangeable licenses. ### [MAI-Code-1.1-Flash](https://headlessthrones.com/models/microsoft-mai-code-1-1-flash) Open weights · 256,000 context tokens ### [MAI-Code-1-Flash](https://headlessthrones.com/models/microsoft-mai-code-1-flash) GitHub Copilot ### [MAI-1-preview](https://headlessthrones.com/models/microsoft-mai-1-preview) API ### [MAI-DS-R1](https://headlessthrones.com/models/microsoft-mai-ds-r1) Open weights ### [MAI-Thinking-1](https://headlessthrones.com/models/microsoft-mai-thinking-1) API