"When I saw the GPT-5.5 evals come back, I literally did a double take."
Yes, we’re model magicians.
It’s exciting to see how Box is pointing GPT-5.5 towards their most difficult, complex knowledge work cases across finance.
A standout moment for Yash’s team:
- GPT-5.5 is amazing at multi-step reasoning and it handled a financial model projection based on structured and unstructured data that resulted in a 19% uplift v. GPT-5.4 Build with GPT-5.5 ➡️ https://lnkd.in/g7KQNhiG
