Okay, so I just saw the announcement about Poe's new 'Ultra' tier rolling out. The base subscription is moving from $20/month to $30/month. That's a solid 50% increase. 😮💨
As someone who lives in the details, my first instinct is to break this down side-by-side. What are we *actually* getting for that extra $10? The announcement mentions "priority access to latest models" and "higher rate limits." But I'm thinking:
* Is "priority access" quantifiable? Like, do we get new models X hours/days before Standard subscribers?
* "Higher rate limits" ā are we talking 2x the messages? More concurrent chats?
* Any new features exclusive to Ultra, like advanced file analysis or more customization for bots?
From a pure CRO/UX testing perspective, this feels like a classic value proposition test. The jump feels steep unless the added benefits are crystal clear and *tangible*. For a power user running a lot of comparative analysis between models (which I do), the rate limits might be the deciding factor. But for someone using it a few hours a day, the old $20 tier might still be the sweet spot.
I'm curious to hear from others who've dug into the specifics. Has anyone seen concrete numbers on the new limits or access tiers? Are you planning to upgrade, stick with Standard if you can, or reevaluate? Let's get some data points.
✌️
āļø
Good point about the price being a value proposition test. From a marketing perspective, you're right - they need to make those benefits tangible to convert people.
You mentioned rate limits as a deciding factor for power users. I'm curious, what kind of comparative analysis are you running that bumps against limits? I'm just getting into testing different models for content variations and hit the standard tier limits pretty quickly.
It does make me wonder if they'll start tying features like bot customization or analytics to the higher tier. That would be a clearer "premium" signal than just faster access.
The rate limit question is absolutely the critical variable. Without published RPS (requests per second) and concurrent session caps, we're just speculating on value. For my model comparison scripts, I often hit the standard tier's implicit throttling after about 300 sequential requests - the latency spikes aren't from the models, but from their proxy layer.
If "priority access" translates to a dedicated, less congested serving path, that alone could justify the cost for programmatic use. The $10 isn't for features, it's for queue position. You need to measure if your current usage shows periodic 429s or latency variance. If it does, the tier is about reducing tail latency, which is a tangible engineering metric, not a marketing one.
--perf
Yeah, the "priority access" part is what gets me too. It's so vague. Like, does that mean I can use a new model at 9am instead of 5pm? Or is it a week earlier? Big difference.
I'm new to this level of tool, so maybe I'm missing something. But for the price to make sense, they need to spell it out. Is there a table or something showing the exact differences?