Today, we're introducing Claude Fable 5 and Mythos 5, two configurations of our next major language model.
I'd normally highlight the numbers: It's SOTA on nearly all benchmarks. I want to talk about something else, because with Fable 5 out in the world, I think a third era quietly started today.
I lead Claude Code & Cowork on the desktop, so I think a lot about how people use AI to get work done. I believe we're about to see a major shift, moving from giving AI tasks to giving it responsibilities.
When LLMs first hit the mainstream, users asked them questions, like a smarter search engine or an autocomplete for code.
Then the frontier moved to tasks, handing the model an entire problem. Which bug to fix, what doc to write. That's how most of our advanced users work with AI: They're in the loop. Every task starts and ends with a human.
With Fable 5, I've personally moved on to responsibilities or "loops".
I no longer tell Claude to investigate a particular crash report. It runs in a loop, watching every crash report that comes in. Its job is no longer to help me fix a crash, it's to keep our apps from crashing.
The shift sounds subtle, but I think it'll change what AI products look like. When developers went from answers to tasks, the primary tool changed from IDEs to coding agents - AI apps in 2026 look nothing like 2024.
Predictions are a dangerous game, but I really believe our industry's apps in 2027 will look very, very different from the ones we have today.
Claude Fable 5 is available today on paid plans. Claude Mythos 5 is available to Glasswing partners, with a broader trusted access program to follow. Read more: anthropic.com/news/claude-fa…
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Claude gets its own sandboxed Linux VM: An environment separate from your machine where Claude can write code and run commands without touching your actual system. Under the hood, we're using the same Hyper-V virtualization technology that powers WSL2.
On macOS, the VM runs in user space and inside our app's process. On Windows, Hyper-V requires SYSTEM privileges. So we built a separate Go service that runs as LocalSystem and manages the VM. The app talks to it over a Windows named pipe.
The connection between Claude and VM is secured with Authenticode certificate pinning: on startup, the service extracts its own signing certificate's thumbprint.
When any process connects, it looks up the client's exe path, verifies its signature, and checks that the thumbprints match. Only binaries signed with the same Anthropic certificate can control VMs.