Our VC-backed subscription days are numbered
Well, the time it takes to compress frontier intelligence down to DeepSeek V4.1 Flash costs (basically too cheap to meter) is dropping, and the differential between the two is also dropping...
So... who cares?
Lastly, I'd like to actually use it in the real world to see how far my plan goes or if its unusable.
Tibo said that the existing $200 subscriptions keep the 20x factor for a while.
Ultrafast would have been nice with the temporary "Pro 400" plan.
But $200 is likely the ceiling of what people will pay for a subscription with usage based on vibes.
$500 for the old $200 is definitely a fumble.
Unless you're talking about buying enough hardware to run something like GLM 5.3, in which case the math just doesn't pencil out—the break even point is several years, and you're stuck with hardware that will be outdated well before then.
There are plenty of good reasons to use local models, but none of them are financial, at least for the vast majority of users.
The optimal move is to retain the minimal access to SOTA models on the $20 plan, and for anything your local model fails at, use SOTA as the backup for either planning or debugging.
This way you're not actually at any disadvantage in terms of capability. You also don't need an advantage, you need to complete the tasks you care about. Eyes on the prize.
RTX 3090 came out a long time ago and it may be 'outdated' at this point but still banging like a champ for anyone who bought one and becoming increasingly more capable as new models unlock it's potential. Hardware hasn't changed much, but what it can do certainly has.
I completely agree about SOTA, but it's a big leap from "you don't need Fable" to "you can get everything done with local Qwen". As always, it depends. Most LLM users are better off with a subscription (or even API pricing) because they won't use AI heavily enough for the hardware to pay off. Then there's the power users who benefit from larger models (software devs, for example). You can argue that there's a middle ground that would do just fine with local models, but I think this group is vanishingly small.
> The optimal move is to retain the minimal access to SOTA models on the $20 plan, and for anything your local model fails at, use SOTA as the backup for either planning or debugging.
Optimal in what way? If I'm having to run tasks twice because the local model effed it up the first time and I'm resorting to my SOTA "backup", that's a waste of my time and far from optimal.
This is missing an important context. And I actually remember this well, because I was saying that too. And the reason I was saying is that $200 plan didn't come with API usage, it was a chat plan.
It made no sense up until they started including API usage. Just as $500 makes no sense now.
> costs going to 10% of white collar income.
There's a permanent and ever lowering ceiling maintained by open weight models. It makes no sense to justify paying 10% of income permanently for something that will get you unlimited local inference for a 6 month subscription cost.
Now you can use it in coding harnesses that call the API.
Why? You can use in codex, right?
Long term, this only works if you have a non-commodity, and if the higher tier is actually more profitable. We'll eventually learn whether both are true. For OpenAI right now, it's probably enough to just increase revenue, even if the higher tier is even less profitable.
Now, as it's linear, it makes much more sense to downgrade to 100$ OAI and pick up a 100$ Claude sub. (without doing the numbers) the usage should remain the same, total paid the same, but having access to best of both worlds. It should be a win for the user, and a loss for OAI.
With this in mind, it sounds like a fumble by OAI.
The vast majority of their revenue comes from large businesses buying for their teams, which are almost certainly not going to juggle lower tiers of different subscriptions to save a few bucks.
Being grandfathered by OAI and happy is not the same as having both, and noticing "hmm maybe Claude is much better for my case, Ill suggest that to our manager"
That they are expensive and climbing doesn't negate my point if the cost of the subscription over how long you plan to keep it is equally or more expensive than the GPUs. You can put together dual 5060 Ti or 5070 Ti systems to run local LLMs too. You don't need to splurge on a 5090. That's a bad option at this point.
I've messed around with Qwen3.6-27B but I'm not sure if it could yet even replace Luna for me.
Qwen3.8-Flash-Next is better still if you can run fast enough. If you have a dual R9700 setup you certainly can. That model is even better.
Qwen4-27B has been announced but not released yet. I'm super pumped for it because I already use 3.8 as my daily driver at home for all my personal stuff, so I'm definitely happy to take an increase in capability.
There is clearly still room for improvement in local models on consumer hardware. With the Qwen 27B models, If you have at least a 5070 Ti I think you can get away with running a small Q4 quant if you use KV cache streaming. The 24GB cards can run Q4 comfortably. If you have a 32B card you can run Q6 comfortably. If you have 48GB ~ 64GB of VRAM you can Q8 comfortably. Using llama.cpp Vulkan let's you pool VRAM across cards (even AMD and NVIDIA etc), so my machine has a 5060 Ti and an R9700.
A dual R9700 rig is really the sweet spot right now with the vLLM-radiance fork. If you can swing a 5070 Ti in there as well to retain some CUDA access, then all the better. That's basically the equivalent to spending 2 years on a subscription, but gets you a system that can run Qwen-3.8-Flash-Next and of course the even more capable Qwen4-Flash when it releases. At the end of the two years it'll run even better models I'm sure.
I'm all in on local now.
I can justify $200/mo but more than double is not appealing to me.
Basically OpenAI aligned with Anthropic on the weekly usage with the caveat that OpenAI doesn't have a 5h limit.
You have to do a lot of things in parallel.
I think this is still true provided you're not using Astra.
An example. Let's assume that the work is evenly divided between days.
Imagine you want to work twice as much.
1. How efficiently can you use the reset credits if they would not reset the normal reset time?
Work with your normal weekly quota 3.5 days, press reset, work with new tokens for the rest of the week. Efficiency 100%.
2. With natural reset time going forward 7 days after each artificial reset.
You work for 3.5 days, press the reset, work for 3.5 days, wait for another 3.5 days for the natural reset, work for 3.5 days, press manual reset, work for 3.5 days, wait for 3.5 days... You can calculate the number for decreased efficiency yourself.
Come on .. this is barely released and you can already make that assessment?
And no, the $200 Anthropic plan is not significantly better than the $200 OpenAI plan, it's just the same Marketing non-sense and anybody shall now rather stick to the $100 plan of both of these provider if the monthly budget is $200. Anthropic doesn't have a Luna Max equivalent, and frankly Sol 6.1 is yet to be thoroughly tested.
Yes, he was talking about safety, but IMHO they're likely already IMHO pushing the boundaries of cartel type behaviour. And they will use safety as the cover to make it happen.
I suspect we'll see serious price fixing and the DOJ do nothing about it because of the inroads these people have with the Trump regime.
Whether that survives contact with Chinese open weight models is hard to say.
Might want to hold off on canceling and continue to bleed them dry until the nerf hits
The only way is for prices to go up. Way up.
The cost of providing the tokens for a heavy user (and let's be frank, the people paying $200 are likely heavy users) is many, many times more than the $200 recurring revenue they generate.
Deepseek has low prices and despite that their profit margin at the beginning of this year was a whooping 82.9%. Since then, they have significantly raised prices.
You can actually check the approx. financials of OpenAI and Anthropic. The growth is insane.
There is no reason to believe why OAI/Anthropic wouldn't have a much better profit margin than DS, taking into account a much higher prices.
> Deepseek has low prices and despite that their profit margin at the beginning of this year was a whooping 82.9%. Since then, they have significantly raised prices.
DeepSeek increased prices substantially not long ago. I find their profit margins hard to inspect considering I have very little idea what sort of environment they may get in China (from cheaper energy to government subsidies). I honestly doubt you have any insight here as well.
> You can actually check the approx. financials of OpenAI and Anthropic.
No you can't. They are not publicly traded, and they constantly and selectively leak bullshit metrics, from extremely unclear ARR, to extremely deceiving EBITDA. You willingly eat their bullshit and call me a picky eater in return.
> There is no reason to believe why OAI/Anthropic wouldn't have a much better profit margin than DS, taking into account a much higher prices.
I see no reason to believe (much less any actual evidence) that OAI or Anthropic have any path to profitability.
If inference (particularly for subscriptions) was in anyway as profitable as you claim today, they wouldn't need private investment rounds like crazy nor they would be desperate to offload this hot potato in an IPO.
82% margins lol. Are you telling me that if you created a machine that turns 1 dollar in 5 what you would do is dillute your ownership of the machine instead of using these fabulous profits to expand the business?
It's clear that you are out of your depth when it comes to financials, business economics or a simple "what it takes to run a business".
You need money to make money. Growth strategy vs. self-financing strategy, pros and cons, when to do each. Critical chain. Limiting factor in infrastructure. Will not expand because this already goes over your head.
I'm not the one out of my depth here.
Feel free to have the last word. I prefer to read idiocy in homeopatic doses.
At the same time, OpenAI is also making its existing $200 Pro plan less appealing. In Codex and Work, $200 Pro subscribers will see their included usage decrease from 20x of what the company offers to Plus users, down to 10x of that same allowance. In ChatGPT, meanwhile, GPT-6 Pro message caps will decrease from 200 to 100 per week.”
https://www.engadget.com/2272106/openai-adds-dollar500-pro-s...
Yikes