Hacker Newsnew | past | comments | ask | show | jobs | submit | maxdo's commentslogin

They just ignore in their benchmarks opus 5 for some reason :) also grok 4.6 . I wonder why

Opus 5 has nerfed cybersecurity performance, making it hard to benchmark: https://support.claude.com/en/articles/14604842-real-time-cy...

Why do you need if every llm vendor has their own harness , what is the moat ?

Spacex even have multi model one


Kimi is expensive . Cursor with subscription is cheaper , grok 4.5 per task paid per tokens ( no subs ) is also cheaper .

If you willing to share to no zdr, meta is waaaaaay cheaper vs Kimi.

With recent offerings from spacex and meta , I hardly imagine why would you pay money to any Chinese vendor it’s not as cheap and it’s not as intelligent neither .

Maybe deepseek is an exception , but it’s only good for narrow use cases that probably goes into modal.com and other gpu + fine tune me easy vendors , not vanilla dumb but cheap model .


Grok is cheaper vs real Chinese frontier aka kimi. Sponsored or not.

What's the cheapest way to use grok models for coding?

Cursor subscription as I mentioned

Cursor ultra is great . For 200 you got essentially unlimited capacity vs Claude.

I used auto in cursor it’s much faster va Claude code and as good.


as a person who spent years building flow.ai before flowise.ai, i'd say drag n drop ui was dead on arrival it works for a set of very hand picked cusomers , the rest were in a weird spot between people who can code( and think accordingly ) and don't.

You hit the nail on the head. A couple years ago I started prototyping a similar node based AI workflow builder which worked well but everyone I showed it to was either; confused at how to use it (even for super basic stuff), or liked it but poked holes by asking "can it do x, can it do y" which it couldn't. There is such a small niche of people technical enough to use something like this that wouldn't just opt in for coding it themselves (especially with AI assisted coding nowadays)

I do still use a part of it on one of my projects for building multi-step image processing pipelines for AI property photo editing (ffocal.io for the curious) that are configurable in the admin panel and it works super well and is useful for modifying and testing pipelines. So there are use cases, they're just very narrow


What I read , my western ego is so big I want to place my opinion everywhere possible .

Image is content , ai can generate art level images and an obvious slop . To read or or not is absolutely up to reader , not someone’s entitled opinion .


Isn't the point of a blog to "place my opinion everywhere possible"? Isn't that why there's an internet?

is kimi that cheap? it's a very expensive model


It's cheaper currently on many of the inference providers.

Personally, I'm having surprisingly good results with DeepSeek 4 Pro at home, which is very good value for money: it's not as good as Claude / GPT 5.6 (I have Co-pilot license at work), but it's still really useful for code reviews, validating thoughts, and especially designing / writing unit tests for new (and old before refactoring) functionality.

And it's very cheap per task. (Flash is even cheaper, but I've had issues with that on more complex tasks where it starts forgetting things and arguing with itself "but wait, let me read the function again").


DeepSeek V4 Pro is ridiculously priced, especially when you take into account caching. According to the DeepSeek usage panel, 50M tokens have cost me $1.38. It's not the smartest and does like to overthink, but if you have well defined problems it's good for coding. Well... except all your data going to China. I just use it for personal projects.


Yup, last month I did ~150mil tokens on DeepSeek v4 Pro for just under $3


Out of interest, are you using the DeepSeek plan?

(I've been using it via OpenRouter and it's much more than that, but still cheap).


I'm not aware of a DeepSeek plan, but I am using the DeepSeek API directly if that's what you mean.


I tried out deepseek v4 pro via a couple providers from openrouter, and it's always getting 429s. Are you running it on your own hardware?


I wish!!

No, I'm using it via OpenRouter in pi.dev - I just used it 30 mins ago... Providers (automatically selected): StreamLake and Baidu Qianfan.


Works fine for me via opencode go.


I toggle back and forth between deepseek v 4 flash/pro on FireWorks.ai using OpenCode. Easy to toggle, I default to flash.


where are you seeing cheap Kimi? pricing I've seen is the same across the board (presumably due to licensing terms) and is in the Terra range.


Cheaper - not cheap!

Morph occasionally have lower prices than the standard rates, and:

https://telnyx.com/pricing/inference-api

Is one which is a bit cheaper... I haven't actually tried K3 myself...


Kimi K3 is fairly cheap per token but thinks like a madman with poor self esteem.


AI slop detected.


why AI revenue will not cover?

Chinese models pushes prices down and quality up, that makes GPU-based automation more affordable, while covering more and more cases to automate.

You can debate that llm producers will go bankrupt, some of them at least for sure.

How do you lose in this market if you do gpu?


> How do you lose in this market if you do gpu?

You don't. Nvidia gets paid either way. They were never the ones in danger (outside of the buildout going bust and having a massive surplus of cheap, used GPUs flood the market).

> You can debate that llm producers will go bankrupt, some of them at least for sure.

And that's the risk that will cascade down and kill off a bunch of companies and cause a debt crisis. If (for example), OpenAI goes to Oracle and says "I promise I'll pay you, at some point in the future, $1T to build my datacenters" and then Oracle funds that build out with debt, and then OpenAI goes bust, or just doesn't make enough money or can't raise enough cash to start making payments on their IOU, Oracle now also can't pay their debt and will eventually go bust, and now the private credit market takes a huge haircut, potentially bankrupting entire funds (like what happened in '08).


I'm sure the bailouts are already getting negotiated


Some LLM companies paid with stocks instead of cash, or with a mix of cash and stocks. Sometimes Nvidia invested in companies and those companies used that money to buy Nvidia GPUs / NPUs.

Nvidia can lose a lot of money if those companies fail. But even in the absolute worst case, Nvidia can go back to making video game GPUs and they will be fine.

The problem is for everyone else who invested in AI. Especially American "401" retirement funds that always invest in the 100 biggest companies.


This is the narrative I completely do not Get. The ai is here , it will eat lots of markets . There’s no way back . If one llm vendor will go bankrupt, their capacity will be picked up. In fact before that it will be absorbed by other vendors . If let’s say OpenAI will fall out of race , it will start slowly on the course of a year selling their commitments to the winner. It’s still unhealthy still tens of billions will evaporate but this is not something can cause apocalypse


if the data center builders cant cover debt payments, market will likely flood with cheap data center gpus to cover some of the liabilities.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: