Comment by baal80spam

Comment by baal80spam a day ago

2 replies

> Use Claude 4.5 Opus

In my org we are experimenting with agentic flows, and we've noticed that model choice matters especially for autonomy.

GPT-5.2 performed much better for long-running tasks. It stayed focused, followed instructions, and completed work more reliably.

Opus 4.5 tended to stop earlier and take shortcuts to hand control back sooner.

8note 18 hours ago

a ralph loop can make claude go til the end, or to a rate limit at least.

opus closes the task and ralph opens it right back up again.

i imagine there's something to the harness for that, too

lukebechtel a day ago

Interesting! Was kinda disappointed with Codex last time I tried it ~2m ago, but things change fast.