MAI-Thinking-1

What the sources document

This entry only states what the sources on enterrsi.si document. Release dates, sizes and other specifications are left out unless a source here gives them.

• The public record for its own MAI models is thin: the MAI-Thinking-1 report describes red teaming but gives no frontier risk classification or government testing. Containment Board, summary
• The MAI-Thinking-1 technical report (June 2026) describes internal red teaming, Microsoft's separate red team and third-party vendors, but names no outside evaluator of dangerous capabilities and no government pre-release testing. Containment Board, outside testing
• The MAI-Thinking-1 report gives safety benchmark results but no frontier risk classification. Containment Board, incident disclosure

Grade context

The Containment Board grades labs, not single models. Microsoft gets C+ overall, rank 4 of 11, on the commitments that decide whether its models can be tested, contained and stopped. See the full report card.

• Safety frameworkB-
• Outside testingC
• Shutdown and pauseB-
• Incident disclosureC
• Whistleblower protectionC+

Sources

1. MAI-Thinking-1: Building a Hill-Climbing Machine Microsoft AI · Jun 2, 2026
2. Frontier Governance Framework (February 2026) Microsoft · Feb 13, 2026