undefined | Better HN

0 pointskelnos1mo ago0 comments

I'm not really sure what to do here. I refuse to give Altman money, but Anthropic keeps disappointing me over and over with crap like this. Gemini seems behind? Not touching Grok.

Meanwhile I've integrated CC into my workflow enough that I'd feel frustrated cutting out all LLM agent use.

I don't have the hardware to run models locally, and I'm not excited about the idea of spending that money. I could use a different harness with one of the services that runs open-weight models for me, but I feel like the cost would be prohibitive. I'm paying $100/mo right now and that's all I'm willing to spend.

0 comments

2 comments · 2 top-level

trymas1mo ago

GLM5.1, Kimi K2.6, MiniMax M2.7

Personally tried GLM subscription. Bought it during new years discount: 36$ for a YEAR.

Cannot burn tokens through with personal project use. From what I can see in stats they allow 25-100M tokens in 5h period (for cheapest plan), depending on the model. GLM5.1 could be a bit slower and likes to (over)think, but I don't see practical differences from Sonnet 4.6 or Opus 4.6.

> I refuse to give Altman money, but Anthropic keeps disappointing me over and over with crap like this. Gemini seems behind? Not touching Grok.

My thought process is totally the same. And even there's slight concern about ethics using GLM, at least in my conciousness, openai is worse and grok is the worst of them all by far, no competition.

InvidFlower1mo ago

I'm not sure if the context limit on the $25/m, and model-size limit on the $100/m would make it not work well enough for OpenCode, but Featherless AI seems a bit unique in terms of how they handle their inference plans.

j / k navigate · click thread line to collapse