Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

$500 GPU outperforms Claude Sonnet on coding benchmarks

github.com

Post reply on HN
Print view View original

281–290 of 311 posts

Re: $500 GPU outperforms Claude Sonnet on coding benchmarks

#281

Post by atoav »

AuthAuth wrote: post #44
atoav wrote: post #28

Earlier quoted context omitted.

It means that the electricity you would have to pay if you did the computations yourself would be more expensive than paying them to do it. Part of thst has to do with the fact that China has cheap electricity, also due to their massive push into renewables. Part of that is just economies of scale. A big server farm can run more efficiently than your PC on average.

cheap electric due to their massive push on non renewables. There has been no change in the price of electricity during the renewable shift.

Normslly you'd expext that more (and cheaper) supply would drive down prices. Classic market logic.

How do do you explain that this market logic ceases to exist for renewables only? A whopping ~2TW or ~35% of generated power in China is renewable and since renewable energy is roughly 1.5 to 4 times cheaper than e.g. coal per kW/h produced that ought to have some impact.

If it has not I'd be curious in your explaination of the mechanism involved.

Re: $500 GPU outperforms Claude Sonnet on coding benchmarks

#283

Post by sznio »

mongrelion wrote: post #171
sznio wrote: post #105

On that topic, anyone here got a decent local coding AI setup for a 12GB VRAM system? I have a Radeon 6700 XT and would like to run autocomplete on it. I can fit some models in the memory and they run quick but are just a tad too dumb. I have 64GB of system ram so I can run larger models and they are at least coherent, but really slow compared to running from VRAM.

Not the answer that you are looking for, but I am a fellow AMD GPU owner, so I want to share my experience. I have a 9070 XT, which has 16GB of VRAM. My understanding from reading around a bunch of forums is that the smallest quant you want to go with is Q4. Below that, the compression starts hurting the results quite a lot, especially for agentic coding. The model might eventually start missing brackets, quotes, etc…

I don't remember exact models, but I tried whatever was available in Ollama. I remember using some really low parameter version of llama

Re: $500 GPU outperforms Claude Sonnet on coding benchmarks

#284

Post by Razengan »

cesarvarela wrote: post #257
Razengan wrote: post #230

Earlier quoted context omitted.

I posted a more detailed report in case you can't see it in your thread view: https://news.ycombinator.com/item?id=47541369 and other comments further back in my history > none of their issues warrant a tantrum on a public forum I don't get frustrated if a problem is genuinely difficult to solve and the product creator is trying their best, I get frustrated when a problem has been solved by other similar products but…

Codex desktop is Electron too. What app are you talking about?

Oh my fuck, I was led to believe that the Codex Mac app was native. Codex itself said so. Google also said so but now it says different.

Codex definitely "feels" more native than Claude: Proper menus etc, like when you right-click on a session in the sidebar, Codex shows an actual context menu, whereas Claude reveals its HTML-rendered jank and highlights the word you right-clicked on as if the sidebar item is just a plain textbox, ugh

Re: $500 GPU outperforms Claude Sonnet on coding benchmarks

#285

Post by dkersten »

miroljub wrote: post #109
mmaunder wrote: post #17

I’d encourage devs to use MiniMax, Kimi, etc for real world tasks that require intelligence. The down sides emerge pretty fast: much higher reasoning token use, slower outputs, and degradation that is palpable. Sadly, you do get what you pay for right now. However that doesn’t prevent you from saving tons through smart model routing, being smart about reasoning budgets, and using max output tokens wisely. And optimiz…

> I’d encourage devs to use MiniMax, Kimi, etc for real world tasks that require intelligence. I use MiniMax daily, mostly for coding tasks, using pi-coding-agent mostly. > The down sides emerge pretty fast: much higher reasoning token use, slower outputs, and degradation that is palpable. I don't care about token use, I pay per request in my cheap coding plan. I didn't notice slower outputs, it's even faster than An…

I’ve also never hit the MiniMax limits and M2.7 is pretty good.

Not as good as Opus, but substantially cheaper!

Re: $500 GPU outperforms Claude Sonnet on coding benchmarks

#286

Post by dkersten »

throwa356262 wrote: post #231
harias wrote: post #146

Earlier quoted context omitted.

MiniMax token plan https://platform.minimax.io/docs/guides/pricing-token-plan

How is the speed and stability? These small Chinese companies dont always have access to serious hardware.

I’ve never had any problems with MiniMax. I wouldn’t call the speed fast exactly, but it’s faster than GLM and seems similar to Opus.

It’s been fast enough that I’ve been using it as my main model (M2.7 and before that, M2.5). Opus still does better at tasks, but MiniMax is so much cheaper. I’ve used their cheaper plan and I’ve never been rate limited.

Re: $500 GPU outperforms Claude Sonnet on coding benchmarks

#287

Post by Epholys »

dahart wrote: post #241
Epholys wrote: post #117

Earlier quoted context omitted.

> The $200 per month subscription comes with a ton of usage. $200 dollars + VAT is half of my rent. I know HN is not a good place to rant on this subject, but I'm often flabbergasted about the number of people here that lives in a bubble with regard to the price of tech. Or just prices in general. I remember someone who said a few years ago (I'm paraphrasing): "You could just use one of the empty room in your house!"…

$200/mo is a lot, sure, but the shocking part of that comparison is your rent. I didn’t know $400/mo apartments still existed. For most people in the US and EU, $200 would be closer to 15%-20% of rent I think? My cell phone bill for my family is almost $200/mo. Last year, at first, $200 seemed crazy. Now that I’m getting addicted to coding agents, not so much. Some companies are paying API rates for AI for employees,…

As you've probably guessed, I don't live in the US, so the price are drastically different. I live in the EU. And for my case, I love in really small flat for some years, so the rent couldn't go up a lot.

> most people in the US and EU, $200 would be closer to 15%-20% of rent I think?

> the average rent is north of $1000/mo.

I really don't know where you get your number from, $1000/mo average is really wild to me. With this amount, you can rent a flat for a whole family in the heart of the city. Nobody of my more well-of friends have a rent this high.

Or maybe you have some capital city in mind like Paris or London?

Re: $500 GPU outperforms Claude Sonnet on coding benchmarks

#288

Post by mlrtime »

bartread wrote: post #247
dakolli wrote: post #244

Earlier quoted context omitted.

must be something incredibly simple you're making out more complicated than it actually is, I've never seen an LLM do these things well.

This is what gives me the warm fuzzies about the HN community: people jumping to wild conclusions about your domain and systems based on a 4 sentence comment. /s

The thing you'll start to notice is this happens A LOT on every subject.

HN tends to think of itself as smarter than the average for every topic. But it turns out there is a lot of bad and factually wrong information in every thread.

Re: $500 GPU outperforms Claude Sonnet on coding benchmarks

#289

Post by Epholys »

hyperbovine wrote: post #274
Epholys wrote: post #152

Earlier quoted context omitted.

I'm not sure I've correctly understood what you're implying. If it's that I'm not working, well, I'm employed. It it's that I'm not working enough to not have this money... Well, we still go back to the bubble. Not everywhere in the world you can easily find a job that pays you enough, even if you accept to work more. And the employer will not accept to give developers a $200/month subscription, even less for persona…

Calm down. I meant that my work covers my pro subscription.

Ha, I misunderstood your comment: in my own language these two words without more context have a very different connotation.

Also, I don't think I was aggressive?

Re: $500 GPU outperforms Claude Sonnet on coding benchmarks

#290

Post by halJordan »

serf wrote: post #236
imiric wrote: post #215

Earlier quoted context omitted.

[flagged]

I have never read a snide comment on this site that i've been more repulsed by. I think because it's so specifically sharpened to stab at the software developer, my compatriot, one of the foremost primary populations here, rather than just an overall shitty human insult -- and timed to do so when the person opens up in an honest dialogue about what they're doing. But good news: every large software house i've talked…

Man I've been waiting for this turn around for years now! This site would lauded whatever it is he wrote for the past 3 years. And it's always been so disheartening to see it accepted just because it was anti ai. Seeing the site get with the times is wonderful
Post reply on HN