Introducing Claude Fable 5

1. 0:00 – 0:15

   Claude Fable 5 launch: most capable public model yet

   1. SPSpeaker0:00

      Today we're launching Claude Fable 5, the most capable model we've ever released to the public. Fable 5 is a Mythos-class model with safeguards that make it ready for general use. We didn't broadly release our previous model with this level of capability because when we finished training

2. 0:15 – 0:31

   Why Mythos Preview wasn’t broadly released: vulnerability discovery risk

   1. SPSpeaker0:15

      and testing it, we saw that the model, Claude Mythos Preview, was finding thousands of cybersecurity vulnerabilities. A model that can find flaws like that can also be used to exploit them. So instead of releasing it, we handed it to the people who protect the world's critical software

3. 0:31 – 0:38

   Redirecting capability toward defense: partnering with critical software protectors

   1. SPSpeaker0:31

      and put it to work fixing the holes before someone could break through them. It was the right call for the moment, but it was never the goal. We

4. 0:38 – 0:45

   Core principle: powerful AI should be safe and accessible

   1. SPSpeaker0:38

      believe powerful AI should be safe and accessible. That's why we went to work on Claude Fable 5.

5. 0:45 – 0:55

   Safeguards as a first-class feature: higher caution than prior models

   1. SPSpeaker0:45

      Every Claude model has safeguards to keep it from doing harm. Fable needed more cautious ones than anything we'd built before. Our safety systems for Fable 5 automatically review

6. 0:55 – 1:05

   Automated review for high-risk domains: cybersecurity and biology

   1. SPSpeaker0:55

      requests that touch on high-risk areas like cybersecurity or biology. Those requests are then redirected to Opus 4.8. We do that intentionally so

7. 1:05 – 1:15

   Routing risky requests to Opus 4.8 to preserve benefits while reducing risk

   1. SPSpeaker1:05

      people can continue to benefit from the capabilities of a powerful model like Fable without the cyber and biology risks that come with it. The safeguards are broad today, but we'll keep

8. 1:15 – 1:18

   Iterating the guardrails: broad now, refined over time to allow safe requests

   1. SPSpeaker1:15

      refining them so that they're better at allowing safe requests.

9. 1:18 – 1:52

   Built for ambitious, long-horizon work: persistence and autonomy

   1. SPSpeaker1:18

      We built Claude Fable 5 for your most ambitious work. It can stay with a problem far longer than any model before it. It's highly autonomous and can operate for days without intervention. And it's not just coding. It can take on projects in finance, research, economics, law, complicated tasks that used to need constant supervision. So point it at something that matters. What's the problem we'll look back on and wonder why it took so long to solve? We know what Claude Fable 5 can do. The interesting part is what you'll do with it. [upbeat music]

Episode duration: 1:53