‹ BackHN Continuity

Thread

Qwen3.8 Max now ranked as the best overall model by agentic index

420 points · 270 comments · apitman

  1. onomojo · · focus · HN ↗
    Any benchmark showing Opus 5 as the best just loses credibility for me. Anyone who's actually used Opus 5 daily knows what I'm talking about.
    1. cromka · · focus · HN ↗
      Agreed, it's extremely frustrating. It's the only model that actually makes me curse when talking to it, even knowing how counterproductive it is.
      1. cromka · · focus · HN ↗
        Replying to myself, because I just bumped into these: <a href="https:&#x2F;&#x2F;www.reddit.com&#x2F;r&#x2F;claude&#x2F;comments&#x2F;1vfvdgz&#x2F;anthropic_losing_aura_slowly&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.reddit.com&#x2F;r&#x2F;claude&#x2F;comments&#x2F;1vfvdgz&#x2F;anthropic_l... <a href="https:&#x2F;&#x2F;www.reddit.com&#x2F;r&#x2F;ClaudeAI&#x2F;comments&#x2F;1vgpyni&#x2F;my_opus_5_experience_in_a_nutshell&#x2F;" rel="nofollow">https:&#x2F;&#x2F;www.reddit.com&#x2F;r&#x2F;ClaudeAI&#x2F;comments&#x2F;1vgpyni&#x2F;my_opus_5...

        Especially the second one seems exactly like my experience.

Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.