How Credits Work on Perspective AI: Visible Cost Per Action
TL;DR: Credits are the single unit of usage on Perspective AI: every chat message, image, video, and agent step draws from one credit balance, and every action shows its cost before you pay it. Credits are displayed in dollar terms, priced at each model's real provider cost plus a thin platform margin, and come from your plan allowance ($14.99/mo Starter includes 200), top-ups, and referrals.
Key Takeaways
- Credits are the single unit for everything on Perspective AI: chat messages, image and video generation, agent operations, and agent previews all draw from one balance.
- Every action shows its cost before your balance moves. Cost legibility is a core product feature, not a settings toggle.
- Credits are displayed in dollar terms, like '$5 of credits', instead of an abstract points system that hides what you spend.
- Each model's rate is its real provider cost plus a thin platform margin, refreshed as provider prices change. There is no per-model markup game.
- Credits come from three places: your monthly plan allowance (200 on the $14.99 Starter, 700 on the $49.99 Pro), top-ups you buy anytime, and referral credits for inviting people.
Quick Answers
How do credits work on Perspective AI?
Every billable action on Perspective AI, whether a chat message, an image or video generation, or an agent operation, draws down credits from a single balance. Text models are metered by tokens sent and received, image and video models per generation, and agents per operation. Every action shows its cost, and the balance is displayed in dollar terms.
How much do credits cost on Perspective AI?
Credits come with your subscription: the $14.99/mo Starter plan includes 200 credits and the $49.99/mo Pro plan includes 700. If you need more, you can buy top-ups at any time, and referral credits add to the same balance.
Why do different models cost different amounts of credits?
Because the underlying providers charge different amounts. Perspective AI prices each model at its real provider cost plus a thin platform margin, so an efficient small model costs a fraction of what a frontier reasoning model or a video generation costs. Rates are refreshed as provider prices change.
Credits are the single unit of usage on Perspective AI: every billable action, whether a chat message to GPT or Claude, an image generation, or a step taken by an agent, draws down one shared credit balance, and every action shows you its cost. There is no separate meter per model, no hidden multiplier per feature, and no abstract points system. You see a dollar-legible balance, you see what each action costs, and you decide what your usage is worth.
That transparency is the actual point of the system. Most AI subscriptions hide consumption behind vague "usage limits" that reset on undocumented schedules. Perspective AI, whose plans are covered in the pricing hub and start at $14.99/mo, treats cost legibility as a core feature: the meter is visible because the company believes you should know what you are spending, not discover it when a limit slams shut mid-project.
One Unit for Everything
Three kinds of activity draw credits on Perspective AI:
- Inference. Messages to any text, reasoning, image, or video model in the catalog.
- Agents. Each operation an agent performs, including tool use and command execution.
- Previews. Trying an agent from the store before you clone it.
Because everything is priced in the same unit, you never have to reason about separate quotas. A week where you mostly chat and a week where you mostly run agents both read as movements of one number, in dollar terms. The balance reads as "$5 of credits", not "4,200 points", so the cost of your AI usage stays comparable to every other line on your budget.
How Each Model Type Is Metered
Metering follows how each model actually bills upstream, rather than flattening everything into a per-message fee that would overcharge light use and undercharge heavy use:
| What you use | How it is metered | What that means in practice |
|---|---|---|
| Text and reasoning models | Tokens: input sent plus output received | Short questions cost little; pasting a 100-page document costs proportionally more |
| Image and video models | Per generation | Each image or clip has a flat, known cost before you generate it |
| Agents | Per operation | Each step, tool call, and command run is metered individually |
Each model carries its own rate. GPT, Claude, Gemini, Grok, and DeepSeek all cost different amounts per token because their providers charge different amounts, and an efficient small model can be an order of magnitude cheaper than a frontier reasoning model for the same prompt. The app shows the cost per action, so routing a simple task to a cheap model becomes a visible, learnable habit rather than guesswork. Over a month, that habit is one of the most concrete ways of saving money on AI subscriptions without giving anything up.
Cost-Plus Pricing, Not a Markup Game
Every model's rate is derived the same way: the real provider cost plus a thin platform margin. Rates are refreshed as provider prices change, in both directions.
This is worth dwelling on, because the aggregator space has an obvious temptation that Perspective AI explicitly rejects: charging opaque per-model premiums on popular models and burying the spread. Under cost-plus pricing, when a provider cuts its prices, the credit rate for that model follows. When a new efficient model undercuts an old one, the saving lands with you, not in a widened margin. The platform makes the same thin margin either way, so it has no incentive to steer you toward expensive models.
The same philosophy explains why there is no artificial "premium tier" of models. As the Perspective AI pricing page details, the $14.99 Starter and $49.99 Pro plans include the identical 60+ model catalog. Plans differ only in how many credits they include, because the honest cost difference between users is volume, not model access.
Where Credits Come From
Your balance is fed from three sources:
- Your plan allowance. Each subscription includes a recurring monthly allowance: 200 credits on the Starter plan at $14.99/mo, 700 credits on the Pro plan at $49.99/mo.
- Top-ups. When a project month outruns your allowance, you buy additional credits on the spot. Top-up credits join the same balance and follow the same per-action pricing, so a heavy month never forces a plan change.
- Referral credits. Inviting people who sign up adds credits to your balance. Referral credits are ordinary credits: same balance, same visible costs, no separate redemption system to learn.
Everything arrives in one pool. There is no expiring bonus currency with different rules, no "promotional credits" that only work on certain models. One balance, one price sheet, visible at all times.
Metered, Not Mined
The credit system also explains something about Perspective AI's business model that flat-rate "unlimited" AI products cannot say cleanly: you are the customer, and the meter proves it.
A flat-rate AI product with hidden limits has two levers when costs bite: quietly degrade the product (throttles, model downgrades, "capacity" errors) or monetize the user some other way. A metered product with visible per-action costs has neither temptation. Revenue tracks usage transparently, so there is no pressure to squeeze value out of your attention or your data to subsidize an unprofitable flat fee. What you pay for is metered, not mined. That is the same design instinct behind the platform's privacy posture, where [private mode](/perspective-ai-private-mode/) keeps chat history browser-only with no server copy, and the transparency page invites you to verify the tracking claims yourself.
How This Compares to Everyone Else's Meter
Every AI subscription meters usage somehow. The difference is honesty about it:
- Single-lab plans (ChatGPT Plus, Claude Pro) sell "usage limits" measured in messages per rolling window, with thresholds that shift without notice. You find the limit by hitting it.
- Budget multi-model apps, including several of the AI tools under $10/mo, advertise low prices that depend on strict caps or heavy routing to cheap models, and the routing is rarely disclosed per message.
- Credit-based aggregators like Poe use points systems, but the points are abstract: a "300 point" message does not tell you what it costs in money.
- Perspective AI shows the cost per action in dollar-legible terms and prices it cost-plus. The meter is the same kind everyone has; it is simply turned to face you.
For the wider market picture, the AI pricing guide compares what every major tool charges and what those prices actually include.
Reading the Meter: A Practical Walkthrough
Here is what the system looks like in an ordinary session. You open a thread and send a short question to a fast, inexpensive model: the response arrives with its cost attached, a small fraction of a credit. You paste in a long report and ask a frontier reasoning model to analyze it: the cost is visibly higher, because tokens in count as much as tokens out, and a 100-page input is most of the bill. You generate two images: each shows its flat per-generation price before and after. You preview an agent from the store for a few steps: each operation meters individually, so trying it costs a known, bounded amount rather than a subscription commitment.
At the end of that session, your balance has moved by an amount you watched happen, denominated in dollars. Nothing needs reconciling at month end, and no invoice surprises you, because the invoice effectively wrote itself in front of you, one action at a time.
Two habits fall out of this quickly. First, input discipline: because text models meter what you send as well as what you get back, trimming a paste-in to the relevant pages visibly halves a cost. Second, model matching: the price gap between an efficient model and a frontier one is on screen, so you stop defaulting to the most expensive model for tasks that do not need it. Neither habit is required, but the meter makes both available, which is more than any hidden-limit subscription can say.
What This Means for How You Use It
In practice, the credit system changes behavior in one direction: toward deliberate model choice. Within a week of seeing per-action costs, most users develop a natural routing instinct: cheap fast models for summaries and quick questions, frontier reasoning models for the work that deserves them, image and video generations made intentionally rather than idly. Your allowance goes further, and you learn what AI actually costs, which is knowledge every other subscription is designed to keep from you.
Try Perspective AI and watch the meter yourself: every model, one balance, and a visible price on every action from your first message.
FAQ
How do credits work on Perspective AI?
Every billable action on Perspective AI, whether a chat message, an image or video generation, or an agent operation, draws down credits from a single balance. Text models are metered by tokens sent and received, image and video models per generation, and agents per operation. Every action shows its cost, and the balance is displayed in dollar terms.
How much do credits cost on Perspective AI?
Credits come with your subscription: the $14.99/mo Starter plan includes 200 credits and the $49.99/mo Pro plan includes 700. If you need more, you can buy top-ups at any time, and referral credits add to the same balance.
Why do different models cost different amounts of credits?
Because the underlying providers charge different amounts. Perspective AI prices each model at its real provider cost plus a thin platform margin, so an efficient small model costs a fraction of what a frontier reasoning model or a video generation costs. Rates are refreshed as provider prices change.
Do unused credits expire?
Your plan allowance renews monthly, and top-up credits join the same balance. For current renewal and rollover terms, check the live credits page in the app, which always reflects the current policy.
Can you earn credits on Perspective AI?
Yes, through the referral program: inviting people who sign up adds referral credits to your balance. Referral credits work exactly like any other credits and draw down against the same visible per-action costs.
Why choose one AI when you can use them all?
Perspective AI gives you ChatGPT, Claude, Gemini, and more in one app.
Try Perspective AI →