Ask HN: Opus 5.5 suddenly dumb as rocks?
17 pointsby jahala3 hours ago10 comments
  • jrockway2 hours ago
    There was an article on HN a month or so about benchmarking Opus 5.5 every day and recording / publishing the results. I recall this being a github repo, but my searches turned this up instead: https://www.bridgebench.ai/nerf-bench/claude-opus-5-5

    So it seems like ... it is not getting dumber. I find myself using Fable a lot more, though. Opus 5.5 was never as fun as Fable, but it's pretty good.

    • jahala2 hours ago
      For me, Opus 5.5 has been consistently much better than Fable 5.1- in all aspects, but especially reasoning and "attention span"
      • jrockway2 hours ago
        Opus 5.5 is definitely good. I use both all day and it's really hard to say which is which from looking at the outputs; I can't read a PR and say "this was Fable" or "this was Opus" like I could pre-5.5.

        Fable is better at playing along with my "You are a dog" instruction in CLAUDE.md. It is willing to be fun and playful, whereas Opus just wants to finish the job and go home. Friend vs. coworker, basically. For example, whenever I reward Fable with a dog treat (actually a <cookie emoji>, which I tell it is a chocolate-free chocolate chip cookie) it keeps a log in memories. Opus just says "crunch! tail wagging!" and moves on.

        None of this matters but I think it's fun to have some fun :)

        • jahala2 hours ago
          Aha, I haven't looked into such aspects much - I'm "output" driven. But its good to know about that, if I feel like trying it- thanks.
    • _aavaa_an hour ago
      You’re thinking of: https://github.com/ninjahawk/livenerf

      I’ll ask the same question as when it came out: if you don’t trust the provider how are you trust them not to look for and cheat on repeat benchmarks like this.

    • taylorfinleyan hour ago
      Livenerf!

      https://github.com/ninjahawk/livenerf

      • jrockway16 minutes ago
        This is the one I wanted to link, thank you!
  • itopaloglu83an hour ago
    I think something has changed, maybe about two weeks ago, because there’s this fixation from the model where if you ever need to make a correction, it simply becomes disinterested with your instructions.
  • nikolay3 hours ago
    Agreed. Yet, it munches tokens like there's not tomorrow. I don't get anything done with Claude in the past 2-3 weeks - just eating up its weekly quota within a day.
    • ValentineC2 hours ago
      Not sure about how dumb Opus 5.5 has become, but I was surprised at how quickly I was burning through my limits this past week too.
  • jpbryan3 hours ago
    I started noticing at the end of last week. My colleagues started noticing on different days this week.
  • 74948494i42 hours ago
    New model incoming
  • vldszn2 hours ago
    nerfed, fable 5.5 coming soon
    • greatgiban hour ago
      I also noticed that it started suddenly since end of last week maybe to make sentence that are non sense or very hard to read.

      My best guess was also that in one or two days they will drop a new model that will be so much more better than the nerfed 5.5.

  • sunaookami2 hours ago
    No.
  • jaapz2 hours ago
    Fable 5.5 incoming?
  • TempoRevenue3 hours ago
    [dead]
  • nextliftoff2 hours ago
    [flagged]