I stopped picking Fable because it refuses based on guardrails so often. I do a lot of security related work, and Fable just won't do any of it. So, I don't bother. Unfortunately, Opus 5 also refuses quite a bit of security work, now, as well, so my Anthropic subscription becomes less useful by the day. Fable may be better, but if it won't do the work...

DeepSeek and Kimi K3 will happily do security work, and they do it pretty well.

▲

vinnymac 5 hours ago | parent [-]

Same, I have gotten good results out of Opus 4.6,4.7,4.8 for my security work though. So I continue to use them for this.

Curious if you’ve find yourself enjoying DS or Kimi more than Opus 4?

	▲	SwellJoe an hour ago \| parent [-]
		DeepSeek is more fun, because it's cheap-as-free, Good Enough, quite fast. I use it for all API stuff, automated runs, testing of the security auditing harness and benchmarks I'm working on, etc. Kimi K3 is smarter, though. At least smarter than DeepSeek V4 Flash 0731. I haven't tried the new Pro version, but will this weekend when I'm working on my personal projects But, K3 has been what I've been using for the actual coding of the harness and such (after Claude models, and then OpenAI models, began refusing to do that work). K3 is very expensive, though. Much more expensive than pretty much everything except Anthropic, and their subscription plans are stingy. I also like Reasonix quite a bit, as an agent harness, though Kimi Code is also very good. I guess I'll try out the new DeepSeek official harness, as well.