OpenAI has published a case study describing how Chatham Financial, a capital markets firm, uses Codex and GPT-5.6 to build technology and redesign workflows. The company reports that trade validation now takes under four minutes, down from 30 minutes previously.
The case study frames the work as both a technology build and a process redesign, rather than a simple model deployment. That distinction matters: the reported gain comes from changing how work is structured, not just from faster inference.
Because this is a single vendor-published source, there are no independent or differing accounts to compare. The numbers and claims are as reported by OpenAI and Chatham Financial, and should be read as such.