So much for DARE program. (TikTok screencap)

  • Paddzr@lemmy.world
    link
    fedilink
    arrow-up
    1
    ·
    12 hours ago

    So how are these paid for AI models vs the free / cheapo models? Who would actually notice the difference?

    • sobchak@programming.dev
      link
      fedilink
      arrow-up
      2
      ·
      7 hours ago

      I’ve been messing around with AI a bit. Kimi k2.6 is very close to Claude Sonnet; “thinks” a lot, but that seems to help it. I think it’s something like 1/3 the cost of Sonnet. The newest GLM is supposed to be similar. Minimax, MiMo, Qwen, and Deepseek is a bit of step down. The big step down are the small models like Qwen 35B and 27B, but they’re still “useful,” and I can run them on my 24GB GPU. I’m also now forced to use Sonnet and crank out slop code :( It’s going to end in disaster.