On: GPT-4.1 in the API

[archeantus]

“GPT‑4.1 scores 54.6% on SWE-bench Verified, improving by 21.4%abs over GPT‑4o and 26.6%abs over GPT‑4.5—making it a leading model for coding.”

4.1 is 26.6% better at coding than 4.5. Got it. Also…see the em dash

[pdabbadabba → archeantus]

What's wrong with the em-dash? That's just...the typographically correct dash AFAIK.

[clbrmbr → pdabbadabba]

Maybe a reference to the OpenAI models loving to output em-dashes?

[drexlspivey → archeantus]

Should have named it 4.10

[clbrmbr → drexlspivey]

But it’s so much weaker than 4.5 in broader tasks… maybe more optimized against benchmarks but it’s just no replacement for a huge model.