No.

I switched to Pi and it was great but I had to switch back to Claude Code purely for the subscription. Even with all the token efficiency I built into Pi it just wasn't nearly as good value as an Anthropic subscription. All open models ended up being more expensive so I switched to using an OpenAI subscription, which was better, but the OpenAI subscription is way less generous AND their models are behind. It became hard to justify.

Sorry, what on earth are you doing where the open models are more expensive than the subs?

When I was using anthropic subs, my "equivalent usage" was $6k/month.

On the open models (specifically deepseek flash), it's < $100/month.

Same work, maybe more.

The large open models (GLM-5.3, Kimi3, etc) are far more expensive than their large closed equivalents (Opus, Sol), because the closed models are far more subsidized when using subscriptions.

If you see no difference at all in between Opus and Deepseek Flash for your work, that is great! Personally, I found the latter to be unusable for the projects I was working on.

Can you describe what you are doing where comparing the two is problematic?

For what it's worth, I use DS flash for statistical analysis, design (C++, Python), scripting, server maintenance. The same I used for Opus.

> When I was using anthropic subs, my "equivalent usage" was $6k/month.

Exactly, yeah. The subscriptions are insanely generous and far better value than open models.

No, my point is I can get the same work done paying < $100/month with the open models.

I don’t think you can.

You'd be surprised what you can accomplish if you don't just let the agents go wild.

An Anthropic sub is $200/mo, max. I only use open weight models, but the reality is that $200/mo Anthropic sub is way better value per dollar than an open weight model. I use open weight models for other reasons, but I have to limit myself to flash, and I am very controlling of which providers I use to keep the price down, and I still come out a little worse than Anthropic for my monthly bill.

I don't know how you use your agents and what kind of work you are doing, but my experience is that in terms of pure output per dollar, you cannot beat an Anthropic subscription with an open weight model.

I'm not saying you don't get good value out of your $100, I'm sure you're doing great, but I'm also absolutely certain if you followed the exact same workflow with an Anthropic subscription you would get better value.

I think the real problem is that if we got rid of money laundering the budget would be even more screwed.

In practice all of these open models are more expensive than a Claude subscription.

Are they? (genuine question, i've got no experience with Claude) This month I used 1 billion tokens (including cache) with DS4.1 flash and it cost me 7% of my $10 opencode go sub.

If you meant in developer time, then I don't know but it seems very fast to me (rarely thinks more than ~10sec, never more than ~30sec) and usually oneshots what I prompt it to do. I use opencode tui as harness.

In my experience yes. Been using Claude since day one and wanted to experiment with Pi for fun, and try to get my costs down. I found the open models more expensive, eventually had to plug my Pi set up into an OpenAI subscription, that was the only way to get good value… Then I realised OpenAI is way behind Opus 5.5 and their subscription is far less generous with usage, so I’ve come full circle and am back on Claude Code.

i found they were until recently. mimo 2.6 flash and glm 5.3 flash seem to finally be cheaper. i find ds4.1 flash still on the high side.

I found GLM Flash competitive price wise but then my tasks took twice as long to complete. Even if it was cheaper (in my experience it still wasn’t) you then lose out on speed because of the insane amounts of inefficient reasoning.

Has he not heard of AI?

They're almost certainly making a profit on subscriptions.

Right, but they are also certainly looking to make higher profits overall by avoiding having to pay for large amounts of (confidential and/or private) training data. Most of what they have would never be for sale either.

Seems smart.

It’s the single biggest problem facing society. If we could stop this rent extraction the economy would prosper.

So astoundingly marginal. Seriously wondering how many Meta astroturfers are in here with some of these posts

What would be great would be a unified car ride app which accessed all vendors. The AI layer adds nothing, I would say detracts actually, and it would not be more popular. UI is just superior to chat in many, many applications.

I think you're just in your own bubble of how people use technology. I hate typing and looking at a screen while in motion. I can absolutely imagine myself using voice mode at home saying "I need a ride to X destination, leaving home within 30 minutes, just me in the car, check Uber Lyft and Waymo and book the cheapest one."

I don't have any kind of agent set up to do that now, but to me that sounds like removed friction even if I use it just a few times a year.

I constantly use the Ask AI thing in Google maps to tell it a complex route involving biking and transit and needing to be at a certain place at a certain time. It hasn't messed up on me yet. I would much rather specify the question in natural language than type anything into a rule based UI. In fact I built a toy app years ago that handled a few of my most common queries with the Maps API and I still never used it because of the friction.

Siri has been able to book you a ride share since iOS 18 without needing an LLM. Generic speech recognition is fine for what you need.

have you used Siri vs. the new crop of personal ai assistants?

siri is horrific

You may think Siri is horrific, but it can do what the poster is asking for at less cost than an LLM.

We’ll see.

Never used Muse or any personal agent for that matter. GTFO.

I just can’t see this ever being common. It sounds terrible. Imagine phoning a number, a person answers and you attempt to relay and negotiate a food delivery order through them, in what world is that better than browsing an app? We don’t have to imagine, this is what we used to do. It sucked and almost everyone stopped doing it in favour of browsing an app. It’s just a superior interface for that sort of thing, it’s already as good as it needs to be.

Different use cases, if the apps don't offer what you need because it's uncommon or custom, then the call is superior, apps can't offer every option, so you default back to messaging and a back and forth, at which point you do want realtime communication

I am not sure what kind of restaurant you're thinking about where it is impossible to put every food option into an app.

If you think about the logistics of running a decent kitchen, the set is quite limited in the first place.

So you need to know exactly what you want to order? People rarely order food like that in my opinion. The majority of the time people like to browse menus and consider the options. In that scenario the voice chat interface is absolutely terrible. “Remind me again of the option you presented to me 3 minutes ago? No not that one, list them again. No I don’t need to know the full item description with all ingredients, just be brief!” and so on. It’s just not an efficient way to relay that type of data.

Funny, I much prefer ordering food by phone call than using apps or websites. I find it much quicker and less error prone.

It doesn’t seem like it to me. Explain?