Why are we still being limited to 25 messages per 3 hours, im paying a lot of money to use this tool , this is absolutely unacceptable. I see people have been complaining for months , what are you doing about it ?
You’re paying $20 / month to get better, faster, more reliable access to
gpt-3.5-turbo and early access to new beta features. You happen to also get 25 messages every 3 hours of GPT-4 access.
That’s up to 200 messages each day or 6,000 messages each month.
On the website responses are limited to ~1,000 tokens, so, with a full context window it’s possible have on the order of 7,000 input tokens and 1,000 output tokens with each message.
For someone doing the same thing in the API that would be ~$0.27 / message.
At 25 messages in three hours that’s $6.75. Eight times a day that’s $54 / day. Over a month that’s up to $1620 in comparative value for what you’d pay using the API.
Then, you get the all-you-can-eat-buffet of
gpt-3.5-turbo on top of that.
So, what, pray tell, do you expect them to “do about it?”
GPUs cost what they cost. Electricity costs what it costs.
You’re getting an absolutely insane amount of value for your money.
A year ago you couldn’t put a price on a product this revolutionary. Even if there were no GPT-4, $20 / month for unlimited GPT-3.5 is a steal.
So, I’m unclear on what, exactly, you want them to do. Do you want them to just give it to you for free?
I mean everyone is entitled to their own opinion, but I would just be embarrassed to express this one publicly myself.
Then they should charge more to lift the cap or increase it, not rocket science. I’m using this tool to change peoples lives, me personally, I need full, unrestricted access and im wiling to pay for it. Can you name many other products that you pay premium service to use that have such limitations ? I cant think of many. If bandwith is the issue, charge me based on usage, i dont care!
They already did that.
You can use as much as you want to pay for, I’m confused what your complaint is.
So… Pay for it.
Does the entire history of cell phone service count?
It’s not bandwidth—the bandwidth used is minimal. It’s the cost of all of the MACC operations done on very expensive compute nodes.
You literally have that option. It’s $0.03 / 1k input tokens and $0.06 / 1k output tokens. All you can eat, go nuts.
There is no conversation.
You’re complaining that you don’t get unlimited GPT-4 access for $20 / month.
It doesn’t matter how much you’re “changing people’s lives” or not.
You say you’re willing to pay for full, unrestricted access.
I told you exactly how to do that.
You’re still completing.
But now you’re also pouting because I don’t believe you are being hampered in your pursuit of making the world a better place because of these troublesome limits—something for which you’ve provided no evidence of for that matter.
Just use the API, pay your bills on time every month, and increase your usage limits as high as you need.
It’s not rocket science.
If you really need unlimited access, it’s right there.
Complaining that a consumer service doesn’t meet your (apparently) enterprise needs is silly.
$20 / month for what you get is insane. If they released this service a year ago—with no free version—people would have been lining up to pay $100 / month for it.