Project
Microsoft's cloud platform for building AI apps, offering a model catalog (OpenAI, Anthropic, others) plus tools for agents, fine-tuning, and orchestration.
Microsoft and NVIDIA say weight openness alone does not determine AI risk, in a joint paper grounded in ISO standards and NIST's AI framework. Microsoft says enterprises should adapt its taxonomy and tie each risk to evaluations, mitigations and an accountable owner.
Reality
Microsoft AI says its 2.8-trillion-parameter MAI-Base-2 will let it license model weights to trusted partners through Foundry. It also says it tunes its models first for its own products and then for Foundry customers, so licensees would come second.
Reality
Azure OpenAI and Foundry in Sweden Central failed for 5 hours 55 minutes on 29 September, according to a postmortem write-up summarised on dev.to. The post prescribes a per-call timeout, a circuit breaker and a second route, and we found two gaps in its sample breaker.
Reality
Microsoft has put Decision-1, a model that scores fixed choices, in Foundry at $0.042 per million input tokens. Agent teams now have to test whether its scores, so far benchmarked only by Microsoft, are calibrated well enough to decide when a case goes to a human.
Perspective Coverage
7 publishersReality
Microsoft released Microsoft-Decision-1, a small model that returns calibrated probabilities for fixed options, priced at $0.042 per million input tokens. The accuracy and speed figures are Microsoft's own, and the base model is due to change soon.
Reality
Anthropic's Claude Code 2.1.285 lets administrators restrict which of eight API providers a machine may use. A company that standardized on one cloud for data residency or committed spend can now hold developers to that choice inside the client.
Reality
Anthropic retires the claude-sonnet-4-5-20250929 model id on November 30, 61 days after its deprecation notice, according to a dev.to walkthrough. Its replacement, claude-sonnet-5-5, costs the same, so the remaining work is finding every key and config that still pins the old string.
Reality
Microsoft released Decision-1, a model that scores fixed answer options and, by its own benchmarks, runs 35 times faster than GPT-6 Sol. Whether routing work moves off general-purpose models now depends on independent tests and on pricing that confirm those self-reported figures.
Reality
Anthropic now includes monthly Claude API credits in Max and Team plans, with Team seats pooled into one balance capped at $500. Whichever Console organization an owner links decides whether that money pays for new internal agents or offsets an API bill the company already pays.
Reality
Greg Brockman said AGI arrived with GPT-6 Astra on September 3. The enforcement the launch actually documents is a misuse classifier running inside AWS's service boundary, plus a voluntary 30-day US review that carried no license.
Perspective Coverage
17 publishersReality
Claude Code v2.1.283 makes auto mode the default, so each shell command gets a Sonnet 5 safety check metered as tokens on Enterprise, API and cloud accounts. Teams on those accounts now pay for command approval unless they choose a different permission mode.
Reality
Microsoft's official plugin lets LiveKit Agents stream live audio to MAI-Transcribe-2-Streaming, a realtime model covering roughly 60 languages. The service is still a public preview that Microsoft does not recommend for production.
Reality
Microsoft's MAI-Transcribe-2 covers 60 languages with speaker labels and word timestamps, though streaming is not among its documented features. Live voice products get a fast model for replies and still need another way to hear the caller.
Perspective Coverage
8 publishersReality
LiveKit Agents now speaks in Microsoft's MAI voices, including MAI-Voice-2-Flash, through an official plugin that covers text-to-speech only. Teams no longer write their own adapter, but every deployment still needs an Azure Speech key and region.
Reality
Cohere's Embed 5 lets teams index with Pro at $0.12 per million text tokens and query with Fast at $0.08 against the same vectors. The quality and throughput figures behind that split come from Cohere's own tests, run on datasets and parsing pipelines a buyer may not share.
Reality
Cohesity's Agent Resilience rolls Bedrock AI agents back to immutable recovery points, yet its product leads say emails already sent stay sent. Outside systems still need their own clean-up and tests, and the company calls its recovery-time edge intended, not measured.
Reality
Microsoft is splitting Copilot into Home, Code and Autopilot, with its always-on Autopilot agent entering private preview in late September. Autopilot agents carry their own identities and bill by usage, so tenant admins need access reviews and spend plans ready before the first instances arrive.
Perspective Coverage
5 publishersReality
Existing subscribers keep their access, but anyone on Free, Go, Plus or the $100 Pro plan is locked out of an upgrade, and a downgrade taken during the pause cannot be undone until OpenAI reopens sales.
Perspective Coverage
4 publishersReality
OpenAI released GPT-6 Astra on September 3, and by September 14 it was generally available on Amazon Bedrock with a million-token input window. Apollo Research had three days with a near-final build.
Perspective Coverage
4 publishersReality
Microsoft says evaluations should decide which GPT-6 model runs each agent. The serving matrix leaves Luna without reserved capacity. GitHub turns both new models on for admins who left the default alone.
Reality
Science · September 21, 2026 · Confirmed2 publishers