A Fairer Way to Manage AI Usage: Cost-Based Budgets, Full Transparency, and Boost Packs
We've fundamentally changed how usage limits work on ARMES — and we think you'll appreciate the reasoning behind it.
Why We Made the Change
Previously, ARMES used quantity-based limits — a set number of queries per day or per week. On the surface, that sounds simple. But it has a fundamental problem: not all queries are created equal.
Asking "What time do the Olympics start and what channel is it on?" counts as one query. Asking an agent to rewrite the Book of Genesis in Gen-Z slang also counts as one query. But these two requests result in vastly different amounts of work for the AI — and vastly different costs. Treating them the same isn't fair to users who ask simple questions, and it isn't sustainable for a platform that gives you access to the world's most powerful AI models.
We needed a system that reflects reality. One that's transparent, fair, and puts you in control.
How It Works Now: Cost-Based Budgets
Every ARMES plan now includes a built-in AI usage budget measured in actual cost — the estimated cost of running AI models in your conversations. More complex requests and more powerful models use more of your budget. Simple questions use very little.
You have two budgets:
- Daily budget — resets every day at midnight UTC
- Monthly budget — resets on the 1st of each month at midnight UTC
These budgets are designed so that the vast majority of users never hit them during normal use. They exist to ensure fair access across the platform and to keep ARMES sustainable as we grow.
Current Budgets by Plan
| Free | Pro Trial | Pro ($19/mo) | Ultra ($200/mo) | |
|---|---|---|---|---|
| Requests / day | 10 | — | — | — |
| Requests / week | 30 | — | — | — |
| Cost budget / day | $0.25 | $1.60 | $3.20 | $20 |
| Cost budget / month | $2.50 | — | $19 | $200 |
Free users retain request-based limits alongside a cost budget as a starter tier. Pro and Ultra subscribers have no request limits at all — only cost budgets, giving you the freedom to use ARMES as much as you want within a transparent, cost-reflective system.
These numbers are subject to change as we monitor and adjust. We've been testing for the past couple of weeks and have landed on something we believe is fair, transparent, and balanced. You can always see the current live limits at armes.ai/ratelimits.
A Note About Pro Trials
The free 3-day Pro Trial gives you full access to all Pro features and agents with a preview budget. This is intentionally lower than the full Pro budget — it's designed to let you explore what Pro offers, not to represent the full experience. When you subscribe, your budgets increase immediately.
Why This Is Better for You
You get every penny of value
Under the old system, a quick factual question consumed the same "query" as a deep research session or a complex multi-step analysis. That meant power users who asked thoughtful, detailed questions burned through limits fast — while simple queries consumed the same allowance despite costing a fraction to run.
With cost-based budgets, lightweight questions barely touch your budget. That means you can ask more questions, have more conversations, and get more value out of your subscription. The cost reflects the actual work, not an arbitrary count.
Different models, different costs
ARMES gives you access to models across the entire spectrum — from ultra-fast lightweight models to the most powerful frontier AI available. Naturally, these cost different amounts to run:
- Lightweight models (e.g., Haiku, Flash) use very little of your budget per message
- Standard models (e.g., Sonnet, Gemini Pro, GPT-5.2) use a moderate amount
- Premium models (e.g., Claude Opus) use more — these are the most capable models in the world
When you use Auto Mode, ARMES automatically selects the right model for your request. A quick question gets a fast, efficient model. A complex analysis gets a more powerful one. Your budget reflects this naturally.
Full transparency
We believe you should always know where you stand. Here's how we've made that real:
- Usage Meter — A new visual meter in your Billing settings shows your daily and monthly usage with color-coded status indicators (green, amber, red) and a countdown to your next reset
- Public rate limits page — Every plan's limits are publicly documented at armes.ai/ratelimits, updated live
- Clear notifications — When you hit a budget, the notification tells you exactly which budget you reached, when it resets, and what your options are
- No hidden math — The Usage Meter shows your usage as a simple percentage. Focus on your work, not your wallet.
Boost Packs: For When You Need More
For Pro and Ultra subscribers who push beyond their included budget, we've built AI Boost Packs — prepaid credit that kicks in automatically so your workflow never hits a hard stop.
How It Works
- Purchase an AI Boost pack (one-time purchase, not a subscription)
- The credit is added to your AI Balance
- If you hit your daily or monthly budget, your AI Balance is used automatically
- You keep chatting without interruption
- While you're within your plan's budget, your AI Balance is never touched
AI Balance only activates as overflow. It sits untouched during normal use and kicks in the moment you need it.
The Packs
| Pack | Price | AI Balance Added | Best For |
|---|---|---|---|
| Boost | $5 | $2.50 | Quick top-up for a focused work session |
| Boost Plus | $15 | $10.00 | Extended heavy usage across models |
| Boost Max | $30 | $25.00 | Sustained power use with premium models |
- No expiration — your AI Balance stays until you use it
- Stackable — buy multiple packs or different packs anytime
- Available to Pro and Ultra subscribers from the Subscription page
Boost Packs exist because we know some users have days or weeks where they push hard — a big project, a research sprint, a product launch. We don't want your AI to stop working when you need it most.
What Happens When You Hit Your Budget
If you reach your daily or monthly budget, you'll temporarily be unable to send new messages. A clear notification tells you:
- Which budget you've reached (daily or monthly)
- Exactly when it resets
- What you can do next (wait for reset, or add AI Balance via a Boost Pack if you're on Pro or Ultra)
Your budget resets automatically — there's nothing you need to do. And if you have AI Balance, you won't even notice the transition.
FAQ
Why did you move away from query-based limits? Because not all queries are equal. A quick factual question and a deep research report are fundamentally different workloads. Cost-based budgets reflect reality and give you more value from every conversation.
Will I use my budget faster with premium models? Yes — more powerful models cost more to run. But Auto Mode is designed to use the right model for the job. Quick questions get efficient models. Complex analysis gets premium ones. Most users find the balance works naturally.
What if I never hit my budget? Then nothing changes for you. The budgets are designed so the vast majority of normal usage stays well within limits. You'll see the Usage Meter in your settings if you ever want to check.
Can I see exactly what I've used? Yes. The Usage Meter in Settings > Billing shows your daily and monthly usage as a visual percentage with color-coded status.
Do Boost Packs expire? No. Your AI Balance stays on your account until you use it.
Can I buy multiple Boost Packs? Yes. They stack. Buy the same pack multiple times or mix different packs.
Where can I see the current limits? armes.ai/ratelimits — always up to date.
The Principle
The goal has always been the same: give you access to the world's best AI, with private inference, at a price that's fair and sustainable for everyone. Cost-based budgets are how we make that real — transparent, proportional, and designed so you get every penny of value out of the AI you're paying for.
We'll continue monitoring and adjusting as the platform grows. If you have feedback on the budgets or the Boost system, we'd love to hear it: joseph@armes.ai.
Joseph Founder, ARMES