News for hackers. Stuff that matters.

A comment on GPT-5.3-Codex. Parent comment

Re:GPT-5.3-Codex by mrandish on Thursday February 05, 2026 @02:34PM (#46903992)
> Yeah, these benchmarks are bogus.

It's not just over-fitting to leading benchmarks, there's also too many degrees of freedom in how a model is tested (harness, etc). Until there's standardized documentation enabling independent replication, it's all just benchmarketing .

[ Parent | on HN ]
Re:GPT-5.3-Codex by fooker on Thursday February 05, 2026 @02:48PM (#46904175)
For the current state of AI, the harness is unfortunately part of the secret sauce.
[ Parent | on HN ]
Re:GPT-5.3-Codex by ndriscoll on Friday February 06, 2026 @10:07AM (#46913715)
In what sense? Codex CLI is FOSS and works fine with other models as a backend, including those served by llama.cpp.
[ Parent | on HN ]