alphabettsy

Isn’t Kimi k3 more expensive than Opus 5.5 in the benchmarks?

Of course, benchmarks don’t tell you how much it will cost on your workload.

GLM 5.3 was more expensive than Haiku and Sonnet 5.5 on a workload I was testing recently despite being much cheaper on the price sheet.

I would love for this to be true, but this is nonsense. The local models that you can run for those prices are not as capable as current frontier. At $200 a month, it would take well over a year to break even, and that’s if you don't include energy prices and time spent tinkering to get and keep it working.

On some benchmarks models like Qwen 3.8 Max which cost < $6/m out cost more than Astra 6 to run at $50/m out. That’s a huge price gap and yet Astra would be cheaper if your work looks like the benchmark.

They already have open models.

https://cursor.com/docs/models-and-pricing

It might be objectively true for you.

Opus 5 is fine for me and works better and faster on low and medium than higher effort on prior versions. Same as 5.6 Sol compared to 5.5 or 5.4.

A total measured over the entire US economy without adjusting for a variety of factors including inflation and population change seems like exactly the kind of thing you would expect from a news headline. How about average or per capita?

As always, it requires evaluation with your work because I’m often finding grok to be much more expensive than the price would lead you to believe.

There’s also the frustration of it not quite being enough sometimes. It’s extremely capable, but I still find that it needs more concrete guidance and boundaries than other models.

GitHub has been providing macOS runners for a while now.

That’s unfortunate. I’m not interested in using Musk associated products anymore than I have to.

I’m trying to understand how this is useful information on its own?

Maybe I missed it, but it doesn’t tell you if it’s more successful for less overall cost?

I can easily make Sonnet 4.6 cost way more than any Opus model because while it’s cheaper per prompt it might take 10x more rounds (or never) solve a problem.

The docs suggest not using max effort in most cases to avoid overthinking :shrug:

That’s been there for awhile.

IMO it doesn’t handily beat it.

It’s typically equivalent, sometimes better, sometimes behind. Better at following a well defined plan, less good at concept exploration and planning imo.

At 1m context it’s basically the same price.

Prompt cache expired?

People keep saying this, but I’m not sure I buy it.

I was using both Codex and Claude Code heavily on some projects this weekend.

In one project Codex was screwing everything up and in another one absolutely killing it. I’ve seen the same from Claude.

In the bad Codex example it had the wrong idea and kept trying to figure out how to accomplish the same thing no matter how many times I attempt to correct it. Undoing the recent changes where it went down the wrong path was the only way to get things back on track.

I wonder if context poisoning is a bigger problem than people realize.

This is my experience too, and I always find these posts confusing. I consider myself a very heavy user 4-6 hrs a day and I never hit limits. I have on the $20 plan but not with Max.

Remind me who was president in 2020..

I believe they make satellite components not consumer hardware in the US

They claim it’s faster and it seems to be so for me.

It’s not an issue with limitations of current technology. In some cases it’s just greed and laziness. I’ve had two vehicles that have the ability to be more friendly to other drivers, but that functionality is only enabled outside of the U.S. (matrix headlights or the equivalent).

GM vehicles had been notorious for having poorly adjusted headlights from the factory. The fact that Xenon systems seemed to always come with auto leveling and LED often does not is crazy.

AI slop.

They’re fantastic imo. Don’t connect to them to the internet.

Then put the warehouse on the first floor and put the store on top.

Lots of big cities have grocery stores with parking garages under them, doesn’t seem much different.

You should still give it a try. Can’t speak for their experience, but doesn’t ring true for me.

The man was prosecuted and the jury found him not guilty.

I would be curious to know what their reasoning was.

I think mine would be that since these agents seem to face absolutely no consequences for their far more egregious actions. Why should something so minor result in jail time. To be clear, I’m not talking about their lawful actions. I’m talking about them assaulting and abusing people and excessive use of force that’s been captured on camera all across the country.

The agents statement was also apparently a lie. Photos showed the sandwich was still in its wrapper on the ground after according to other articles like this one https://reason.com/2025/11/06/d-c-jury-acquits-sandwich-guy-...

Fair question. The jury in this case decided it wasn’t assault.

My hypothesis is that people generally feel that police face little to no accountability and so there is a more serious double standard to contend with.

Are you sure you’re not conflating knowledge with intelligence?

True, but more and wider references seems to imply better when I’m not sure that’s true. Wikipedia is edited and it’s sources are curated. I think that’s a good thing.

This keeps happening.

2022: https://www.xda-developers.com/google-pixel-3-update-emergen...