Credits metered from real model cost

Priced by what
each run costs.

Each run is billed from what its model calls cost: one credit per $0.0005 of model spend, about 2 credits on an economy model and 15 on a standard one. Pick a plan that matches your usage; bring more credits when you need them.

Paid plans are not currently available in the United Kingdom.

Free

$0/mo

25 Credits/mo

≈ 12 runs on economy models

Try the optimizer with our cheapest models — no card required.

25 credits per month
Economy models — fast, current, ~2 credits a run
5 requests/minute, 20/day
History tracking (7 days)
Community support

Starter

$3/mo

600 Credits/mo

≈ 40 runs on standard models

A taste of the full catalog without committing to a real plan.

600 credits per month
Access to ALL models (Gemini, Claude, GPT, OpenRouter)
10 requests/minute, 200/day
Batch refinement (up to 5 prompts)
Most Popular

Plus

$5/mo

1,200 Credits/mo

≈ 80 runs on standard models

For hobbyists running prompts every day — the sweet spot under $10.

1,200 credits per month
All models including Claude Sonnet & GPT-5
15 requests/minute, 500/day
Batch refinement (up to 10 prompts)
Unlimited history export (JSON/CSV)

Pro

$10/mo

2,800 Credits/mo

≈ 180 runs on standard models

For pros who want headroom without paying business prices.

2,800 credits per month
All models — including Claude Opus runs
30 requests/minute, 1,500/day
Batch refinement (up to 20 prompts)
Priority processing
Credit top-ups

Need more credits this month?

One-time top-up packs never expire. Add on top of any subscription, including Free. Use them for the occasional Opus run without upgrading your whole plan.

$5
1,000 credits
$20~12% off
4,500 credits
$50~17% off
12,000 credits
Transparent billing

How Credits Work

Every optimization calls real AI models, and each run is billed from what those calls cost upstream: one credit for every $0.0005, rounded up, one credit minimum. Premium models such as Claude Opus cost more per run — the credit calculator prices each one.

Your plan price buys the month's credits and the pipeline that spends them — classification, framework selection, expansion and scoring — so on monthly billing a credit costs you $0.0036 (Pro) to $0.0075 (Business), not $0.0005.

~2
Credits · economy run
~15
Credits · standard run

Average charge, production runs Aug–Oct 2026

Feature gates per tier

Optional OpenRouter parameters unlock as you move up tiers. Backend rejects requests that exceed your plan; the UI dims the matching control.

Web search

OpenRouter web plugin (Exa) augments the prompt with live results.

Free
Starter
Plus
Pro
Business
Fallback chain

Pin up to 3 alternate models per request — first available wins.

Free
Starter
Plus
Pro
Business
Advanced reasoning depth

Reasoning effort high / xhigh on supported models.

Free
Starter
Plus
Pro
Business
Zero Data Retention

Routes only to endpoints that don't store your prompt.

Free
Starter
Plus
Pro
Business
Premium models

Opus, GPT-5, o1/o3, Gemini Pro, Grok-4 — flagship reasoning tier.

Free
Starter
Plus
Pro
Business
Batch size (prompts/job)
Free2
Starter5
Plus10
Pro20
Business30

How credits are reserved

Each optimization holds credits upfront based on the chosen model’s cost class. The hold is not the price: after the run it is settled to the measured cost, so what a run actually charges is the figure under each class.

Economy1 credit held

Gemini Flash Lite · other economy-band models

A run averages ~2 credits

Standard5 credits held

Gemini Flash · Claude Haiku · GPT mini · DeepSeek

A run averages ~15 credits

Premium30 credits held

Opus · Sonnet · GPT-5 · Gemini Pro · Grok-4

A run is priced per model

Reconciliation: if the actual upstream cost is lower than the reservation, the excess is returned to your subscription bucket as a reservation_refund ledger row. If higher, the difference is charged from your balance.

Pricing questions, answered

How credits, top-ups, and per-seat billing actually behave.

What is a credit on VantagePrompt?
A credit measures model cost, not a flat fee: each run is billed one credit for every $0.0005 its model calls cost upstream, rounded up, with a one-credit minimum. Measured on production runs, an economy-model run averages about 2 credits and a standard-model run (Gemini Flash, Claude Haiku, GPT mini) about 15. Premium models such as Claude Opus cost more per run; the credit calculator prices every model. The plan price buys those credits and the pipeline around each model call, so on monthly billing a credit costs $0.0036 (Pro) to $0.0075 (Business).
Do unused credits roll over to the next month?
Subscription credits reset to your plan amount on the 1st of each month and do not roll over. Top-up credits are separate: they never expire and carry over indefinitely. Runs always drain the subscription balance first, so purchased top-ups are only touched once the monthly allowance is gone.
What happens if I run out of credits mid-month?
Buy a one-time top-up pack: 1,000 credits for $5, 4,500 for $20, or 12,000 for $50. Packs never expire and stack on top of any plan, including Free. You can also enable auto top-up, which charges your saved pack automatically when the balance runs low.
Which AI models can I use on each plan?
The Free plan uses the economy band — fast, current models that average about 2 credits a run. Every paid plan, starting with Starter at $3/month, unlocks the full curated catalog: Gemini, Claude, GPT and the rest of the OpenRouter whitelist, including premium reasoning models.
How does Business per-seat pricing work?
The Business plan costs $15 per seat per month with a minimum of 2 seats. Each seat adds 2,000 credits per month to a shared team pool, and admins can set per-member quotas on top. The prompt history is shared across the whole workspace.
Can I cancel or change my plan at any time?
Yes. You can upgrade, downgrade, or cancel from the billing portal at any time, and your plan stays active until the end of the period you already paid for. Annual billing, where available, costs ten times the monthly price — two months free.

Need enterprise-scale volume?