I asked Claude Fable 5.5 to explain OpenAI's new 23-page proof about π. It came back with a narrated 2-minute 3D animation that walks through the proof itself. The claim: the irrationality exponent of π is EXACTLY 2. Since 1953 the best proven bound had only crawled from 42 down to about 7.

I asked Claude Fable 5.5 to explain OpenAI's new 23-page proof on the free group factor problem. It came back with a 2-minute 3D animation. The claim: a question open for over half a century, answered. L(F2) and L(F3) are the SAME algebra. Lean proof included. Watch:

Okay, my new 23-task eval suite for @VulcanBench is done. A lot of little details I wanted to get right with this. It was important to me that I had more broad language coverage, and also that it takes into account the issues that OpenAI raised yesterday with SWE-Bench Pro's four failure modes. I'm new to benchmarking, but learning more every time I create a new set of evals. I'm calling this set v3, getting ready to run the first smoke tests. Then this weekend, phew, safe to say I have a lot of models to test! I will be committing these to Github so anyone can take a look and give me feedback on these as well. Live long and Benchmark 🖖