That seems really high! Higher than weâre seeing on our end. I sent you a separate email to try to get to the bottom of things.
This change is not sustainable for me at all. I have already run out of credits within 6-7 days into my pro plan while before the credits covered the month easily for my users and I have less users than before. Unfortunately I will have to switch to custom GPTs and poe as Iâll see a $150-$200 USD additional per month maybe higher. And scaling these tools just amplifies the problem of usage.
Thatâs a great solution! Pickaxe is awesome when you need to directly monetize your tools, but model cost can be a concern. If youâd like to continue to be able to use Pickaxe, Iâd take a look at our new model cost estimates in the builder. Some models are more than a THOUSAND TIMES cheaper than others⌠so cost issues could be resolved quite quickly by making a switch. Let us know if we can help!
For example, with a Pro subscription you get $50/m of usage.
With OpenAI o1, thatâd get you only about 200 interactions
With GPT-5, thatâd get you about 5K interactions
With Gemini 2.0 Flash, thatâd get you almost a MILLION uses.
Just sharing this $10 example for perspective. Choosing the right model can make a huge difference in how long your credits last. Some models are made for high-volume simple tasks while others handle complex reasoning and naturally use more credits per message. Pro users get 5 times this usage included at no extra cost.
| Model | Approx. Messages for $10 | Task Type | Example Use Cases |
|---|---|---|---|
| Google Gemini 2.0 Flash | ~188,679 | Simple & High-Volume | ⢠High-volume, simple Q&A |
| OpenAI GPT-5 Nano | ~33,333 | Simple & High-Volume | ⢠Basic customer support bots |
| Claude Haiku 4.5 | ~4,000 | Simple & High-Volume | ⢠Quick data classification |
| Grok | ~358 | Complex & High-Stakes | ⢠Balanced performance for general tasks |
| Gemini 2.5 Pro Preview | ~191 | Complex & High-Stakes | ⢠Advanced coding assistance |
| Claude 4 Opus | ~77 | Complex & High-Stakes | ⢠In-depth content creation |
| OpenAI o1 | ~45 | Complex & High-Stakes | ⢠Complex financial modeling |
Ya I was monetizing the tools but the new credit system is not cheaper at all. Itâs substantially more expensive. Iâm using Claude for all of my tools. In the last 7 days there have been 537 interactions. But over the last 30 days Iâve had 1800 usages and never came close to reaching my cap. So this change has impacted the tools I created.
If Claude is your preferred model, then Haiku 4.5 is the latest and most affordable option to consider. In fact, your total usage would cost less than $1, which is already well below the $50 monthly allowance included in the Pro plan. Claude Haiku 4.5 costs around $0.0025 per 100 interactions, so your 537 interactions in 7 days would come to about $0.013, and even 1,800 interactions in a month would cost roughly $0.045 if Haiku was selected properly. Choosing lighter models like Haiku instead of heavier ones such as Opus or o1 can help your credits last much longer while maintaining solid performance.
Thank you for this insight. Iâll switch my tools to Haiku.
Exactly! Now weâre forwarding the actual costs on to you, so you can actually see how much your usage costs. That might prompt a need to make a model change, but we think itâs better to be transparent, rather than hiding the costs as we were doing before.
Totally get it and appreciate the transparency.
Hello, thank you for explaining everything in the best possible way. I am a LTD user; could you tell me how much monthly credit each Tier 1, 2, and 3 is being given? Thank you!
You should be able to see that for your LTD in your account page here!
Nathaniel, thank you very much, I am Tier 3 and I have $15, I would like to know how much Tier 2 and Tier 1 ended up with, is that information secret?
I believe that the increased transparency of credits and the ability to calculate all tools using API pricing through this recent change represents a truly courageous improvement. I express my heartfelt gratitude for this.
So if BYOK support is going away then what do we tell existing clients? Hell, I donât even have contact with many of them. I just build the studio and move on. What happens when all of these studios suddenly break due to this change?
Iâd be happy to jump on a call with you to discuss this, but Iâm quite confused. If youâve built studios for clients and left them behind, did you transfer the studios to them? Thatâd mean they have their own Pickaxe subscription, so there shouldnât be a problem.
On the other hand, if the studios you built are still managed by your account, are you using your API keys currently or ones provided by clients? If youâre using your own, again, no problem. If youâre using API keys provided by clients, and the studios are owned by you, then youâd have had to reach out anyway, as over time model providers update their key standards anyway.
Itâd help a lot if youâd explain your situation in detail!
Hi @nathaniel,
I can see you changed the rules.. This is not what we paid for anymore. What happens to the â â â â â â â buyers?⌠You canât change the product we bought⌠itâs not right. This is what we paid for: View plan details
Deal terms:
-
Lifetime access to Pickaxe
-
All future Gold Plan updates
-
If Plan name changes, deal will be mapped to the new Plan name with all accompanying updates
-
No codes, no stackingâjust choose the plan thatâs right for you
-
You must activate your license within 60 days of purchase
-
Ability to upgrade between 3 license tiers while the deal is available
-
Ability to downgrade between 3 license tiers within 60 days of purchase
-
Available for new Pickaxe users and returning â â â â â â â purchasers
-
Previous â â â â â â â customers who purchased PIckaxe can upgrade their license to increase their feature limits
-
Previous â â â â â â â customers will be grandfathered into any new features and feature limits
-
Monthly credit limits are based on GPT4 and subject to change based on costs of newer versions
Features included in all plans:
-
Create unlimited Pickaxes
-
Embed Pickaxes anywhere
-
Custom domain for Studios
-
Set embed usage limits
-
Branded chat icons
-
View how people use your prompt
-
10% take rate
License Tier 3 features:
-
1,000 monthly credits
-
50 studios
-
3 action(s) per tool (API integrations)
-
Bring your own OpenAI key (unlimited usage)
Is there a refund for us?
Hi @siderick,
The deal terms you shared include this line:
âMonthly credit limits are based on GPT4 and subject to change based on costs of newer versions.â
Because of that, Iâm not seeing anything that isnât being honored. It looks more like a misunderstanding of how AI model pricing works today. Early ChatGPT used simple message limits, but now every major model provider, including OpenAI and Gemini, charges by tokens. The new credit system follows the same standard: the more a model is used, the more it costs.
If your goal is to get the most out of your lifetime plan, Iâm happy to help you optimize your usage. And if the concern is about the ethical or moral side of credits, all pricing is public and easy to cross check. OpenAI and other models charge based on API usage, and Pickaxe mirrors that same structure.
Hi! Great question! To follow up on Ahbiâs points above, as you can see, your plan includes âAll future Gold Plan updatesâ.
This is an update that affect all users. You included! But never fear, this is actually a huge improvement for you. From your post, it sounds like economy and getting the most out of Pickaxe is super important to you. In that case, with the new system⌠youâre in luck! With a model like Gemini 2.0 Flash (as you can see here), instead of getting a paltry 1K uses, you can now get over 200 thousand uses! All without paying a cent beyond whatâs included in your plan.
And, if youâre worried about BYOK, donât be! We now allow as needed billing, only required when you go over your included limits. Thatâs even better than what we offered before, because you only need to pay when you exceed your included limit!
Let us know if we can provide any further explanation, and youâre welcome to reach out to your LTD provider for help if you need it!
I would love some help with optimising the usage on my main agent. It is costing disproportionally high regardless of what model is used.
Nathaniel, please, there is a very fine line between being transparent and bullshiting, donât do a âtrumpâ on your clients. Many of us overnight went from zero extra a month to $200-400 extra per month. As long as pickaxe is not overcharging us on credits, thatâs the new reality we need to live in. Instead of bs examples of gemini 2.0 flash, who cares, we are not using it, we need to focus on the root cause. Root cause is that we have not been optimising our agents and they are wasting tokens left right and center. We need to make advanced models cost efficient not settle for old cheap models to save costs. We need to address this together before you start losing clients like flies. I would really appreciate pickaxe guidance on what contributes to costs, how to save, dos and donâts, etc. I would even pay a small fee for someone to optimise my agents for cost without compromising performance.