Hacker News Reader

Given this comment, I tried it.

It's no where close to Claude, and it's also not better than OpenAI.

I'm so confused as to how people judge these things.

I'm confused as to how you haven't found R1 to be much better. My experience has been exactly like that of the OP's

What type of prompts were you feeding it? My limited understanding is that reasoning models will outperform LLMs like GPT-4/Claude at certain tasks but not others. Prompts that have answers that are more fuzzy and less deterministic (ie. soft sciences) will see reasoning models underperform because their training revolves around RL with rewards.

Where are you guys using the full model?

Doesn't it require 220GB ram? I only se V-3 on their website and the distills available to run locally.

You have to pay if you want to use the full API but you can chat with it for free https://www.deepseek.com/

Just try it on their site or via the api

Reply on news.ycombinator.com