Results from Internal Coding Evals For Claude Fable
- For 98% of tasks, it simply does the same thing as GPT 5.5 or Opus 4.8 and costs 2x
- For 2% of hard coding tasks, it does make sense if you are willing to pay double and get some quality gains
So ideally, you want to ROUTE VERY hard tasks to Fable
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.
Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use.
Its capabilities exceed those of any model we’ve ever made generally available.
Stop saying that Opus 4.6 is better than Opus 4.8.
I just tested Claude Opus 4.8 (Right) and Claude Opus 4.6 (Left) on the same exact prompt.
Claude Opus 4.8 completed it in less than 2 minutes.
Claude Opus 4.6 "Garnished" for over 7 minutes.
Stop using Opus 4.6.
We recently submitted a confidential S-1. We expect it to leak so we’re just announcing it. We have not decided on timing yet; it may be a while because there are things we want to do that are likely easier as a private company. But it’s a complicated set of tradeoffs and this gives us the option to go public sooner if that ends up being best.
This announcement is being made pursuant to Rule 135 under the Securities Act of 1933, as amended, and does not constitute an offer to sell or the solicitation of an offer to buy any securities. Any offers, solicitations of offers to buy, or any sales of securities will be made in accordance with the registration requirements of the Securities Act.