Thread

1. ApolloFortyNine · 1790001735 · focus · HN ↗

   The model being tested is 18k as configured.

   I didn't expect this to make the 5090 to look like a good deal.

   1. nacs · 1790007044 · focus · HN ↗

      5090 has 32GB VRAM.

      It'd be silly to buy the 18k model to run a tiny model like Qwen 27B. You use models like GLM Flash and Qwen Next which won't fit on a single 5090.

      1. asimovDev · 1790013265 · focus · HN ↗

         can run multiple subagents of Qwen 27B though, right? Unless I am fundamentally misunderstanding how VRAM constraints work

         1. Eisenstein · 1790014902 · focus · HN ↗

            You might be. Running another agent doesn't load a set of new weights. It creates a new KV cache for the agent and adds the prompts to the queue. Its just another inference turn.

            1. asimovDev · 1790074950 · focus · HN ↗

               thanks, I naively assumed when, for example, Claude Code starts subagents it loads a new instance with empty context

Unofficial Hacker News client; not affiliated with Y Combinator.