Radox AI Pricing and Plans 2026: Six Tiers From $0 to $50, and What Happens When the Tokens Run Out
Radox AI sells tokens, not seats. Six plans (Free, Starter, Basic, Pro, Scale, Unlimited) differ only in monthly token allowance and per-minute limits. This guide works out the price per million tokens, the yearly saving, what happens after the allowance is gone, which payment methods are accepted, and exactly what the refund rules say. This article focuses on chat/text.
| Free plan | 10M tokens/mo, 60 req/min |
| Entry paid plan | Starter, $4.50/mo for 100M |
| Top plan | Unlimited, $50/mo |
| Yearly billing | "2 months free" |
| Allowance used up | Bills balance; empty balance returns HTTP 402 |
A price list that fits on one screen, with two billing rules that matter more than the sticker price.
Every plan reaches the same models. What you buy is a monthly token pool and a per-minute ceiling, and what happens after the pool is empty.
The six plans are easy to compare because only two numbers change: tokens per month and requests per minute. The part worth a closer look is what sits behind the allowance. Unused tokens do not roll over, and once the pool is empty you are billed from a prepaid balance at each model's own per-token rate, which is not the same figure as the plan's per-1M price. I would pick a plan from the allowance arithmetic below, then look up the rate of the one model I actually plan to call.
Six rungs, one ladder: more tokens, higher request ceilings, and one price that has no token cap.
Source: radoxai.com/pricing, Oct 15, 2026. The pricing page states: "The only difference is how many tokens you get each month."
- A monthly token allowance that resets each month
- A requests-per-minute and tokens-per-minute ceiling
- Support tier: community, email, priority or dedicated
- Rollover of unused tokens
- Any extra model access
- A fixed price for overage: that is per model
Four numbers that frame the whole price list.
Starter, 100M tokens. Pricing page.
Unlimited, no monthly token cap.
No card required, 60 requests/min.
Yearly toggle on the pricing page.
| Number on the page | What it tells you | What it does not tell you |
|---|---|---|
| Tokens per month | Size of the included pool | How many requests that is for your prompts |
| Requests per minute | Ceiling on bursts of calls | Whether your workload ever reaches it |
| Per 1M figure | Price of the included tokens | Price per token after the pool is empty |
No independent customer reviews of the billing found, so these are labeled vendor statements.
We searched public review and community sources on Oct 15, 2026 and found no independent customer accounts of Radox AI billing, so none are summarized. The cards below quote Radox's own pricing page and Terms so you can see which billing promises are written down.
"With an empty balance you get a clear 402 instead of a surprise invoice."
"It resets each billing period and unused tokens do not carry over."
Rates can change; "we will announce increases in advance, and a change never applies retroactively to usage already billed."
Each step up gives you more tokens and a lower price per million.
Free
10M tokens / month · 60 req/min · All chat models · Community support.
Starter
100M tokens · $0.045 per 1M · 120 req/min · Email support. Yearly $45.
Basic
200M tokens · $0.0425 per 1M · 300 req/min · Usage analytics · Email support. Yearly $85.
Pro
350M tokens · $0.0357 per 1M · 600 req/min · Usage analytics · Priority support. Yearly $125.
Scale
500M tokens · $0.031 per 1M · 1,200 req/min · early access · Priority support. Yearly $155.
Unlimited
Unlimited tokens · 3,000 req/min · early access · Dedicated support. Yearly $500.
Source: radoxai.com/pricing, Oct 15, 2026. "Best value" is Radox's own label on the Pro card. On Unlimited the page says: "No monthly token cap. Rate limits still apply."
Swipe sideways to see every column.
| Plan | Price / mo | Tokens / mo | Per 1M | Requests / min | Tokens / min |
|---|---|---|---|---|---|
| Free | Free | 10M | n/a | 60 | 100,000 |
| Starter | $4.50 | 100M | $0.045 | 120 | 200,000 |
| Basic | $8.50 | 200M | $0.0425 | 300 | 500,000 |
| Pro | $12.50 | 350M | $0.0357 | 600 | 1,000,000 |
| Scale | $15.50 | 500M | $0.031 | 1,200 | 2,000,000 |
| Unlimited | $50 | No cap | n/a | 3,000 | 5,000,000 |
Swipe sideways to see every column.
| Plan | Support | Usage analytics | Early access to models | Custom limits |
|---|---|---|---|---|
| Free | Community | Not listed | Not listed | Not listed |
| Starter | Not listed | Not listed | Not listed | |
| Basic | Listed | Not listed | Not listed | |
| Pro | Priority | Listed | Not listed | Not listed |
| Scale | Priority | Listed | Listed | Not listed |
| Unlimited | Dedicated | Listed | Listed | On request |
"2 months free" checks out on every paid plan.
The pricing page has a Monthly / Yearly toggle labeled "2 months free". Comparing twelve monthly payments with the yearly price confirms the claim exactly: each yearly price is ten months of the monthly price.
Swipe sideways to see every column.
| Plan | 12 x monthly | Yearly price | Saving | Shown per month |
|---|---|---|---|---|
| Starter | $54.00 | $45 | $9.00 | $3.75 |
| Basic | $102.00 | $85 | $17.00 | $7.08 |
| Pro | $150.00 | $125 | $25.00 | $10.42 |
| Scale | $186.00 | $155 | $31.00 | $12.92 |
| Unlimited | $600.00 | $500 | $100.00 | $41.67 |
Mind the commitment. The Terms say "Paid plans are charged in advance for the period you select", and Radox's Refund Policy says the 7-day guarantee "applies only to your first subscription" and "does not apply to renewals or plan changes" unless a platform technical failure prevented use. A yearly payment is therefore a bigger up-front sum with a short refund window. Starting monthly is the lower-risk way to test.
Pick a tier by the tokens you will actually send.
Free: 10M tokens a month and no card. Check how your client behaves first.
Starter ($4.50) or Basic ($8.50). Note the 120 and 300 requests-per-minute ceilings.
Pro ($12.50, 350M, 600 req/min), the plan Radox labels "Best value".
Scale ($15.50, 500M, 1,200 req/min), or Unlimited ($50) if you regularly exceed the pool.
Start on Free or one paid month, read the usage dashboard, then choose. Switching is allowed any time.
Nothing breaks, but the price per token changes.
Per the pricing page: "Once your allowance is used, requests bill against your account balance at each model's per-token rate". The Terms describe two layers: a plan allowance that "resets each billing period and unused tokens do not carry over", and a prepaid account balance that is "charged at each model's published per-token rate once your plan allowance is exhausted". If both are empty, requests are refused with HTTP 402 "before reaching a model", so you are not billed for something you did not authorise.
| Question | Documented answer | Source |
|---|---|---|
| Do unused plan tokens roll over? | No: "unused tokens do not carry over" | Terms, section 4 |
| Does the prepaid balance expire? | "Balance does not expire while your account is open" | Terms, section 4 |
| What is returned when both are empty? | HTTP 402 | Pricing page and Terms |
| Do upgrades apply immediately? | "Upgrades take effect immediately and your allowance is refreshed" | Pricing page FAQ |
| When do downgrades apply? | "Downgrades apply from the next billing period" | Pricing page FAQ |
| Can rates change? | Yes, with advance announcement and no retroactive effect | Terms, section 4 |
How one billing month plays out.
Your plan's token pool starts again. Anything left from last month is gone.
"Requests draw from the allowance first, so you are not billed per call until it runs out."
Balance is charged at each model's rate; an empty balance returns 402 until you top up or upgrade.
Allowance tokens and balance tokens are priced on different lists.
The homepage lists per-model rates. Two examples with equal input and output rates make the arithmetic exact: deepseek-v4-flash-0731 at $0.115 per 1M tokens and gpt-5.6-luna at $0.357 per 1M tokens (homepage model list, Oct 15, 2026). Rates differ by model and can change, so treat these as worked examples, not a quote.
| Scenario | Arithmetic | Result |
|---|---|---|
| Starter's 100M tokens inside the allowance | $4.50 / 100 | $0.045 per 1M |
| Same 100M tokens from balance on deepseek-v4-flash-0731 | 100 x $0.115 | $11.50 |
| Same 100M tokens from balance on gpt-5.6-luna | 100 x $0.357 | $35.70 |
| Free plan's 10M tokens at the luna rate | 10 x $0.357 | $3.57 |
Where Unlimited starts to make sense
Assume you are on Scale ($15.50, 500M included) and every extra token is billed at one model's rate. Unlimited costs $50, so it wins once overage would exceed $34.50.
| Overage model rate | Extra tokens that cost $34.50 | Total monthly tokens at break-even |
|---|---|---|
| $0.115 per 1M (deepseek-v4-flash-0731) | 300M | 800M |
| $0.357 per 1M (gpt-5.6-luna) | about 96.6M | about 596.6M |
Cards and wallets through Paddle, plus Binance Pay and crypto.
| Method | What the pages say |
|---|---|
| Paddle | Credit/Debit cards, Apple Pay, Google Pay, iDEAL |
| Binance Pay | Verified automatically through gateway API integrations or transaction hashes |
| Cryptocurrency | USDT/BTC; submit the correct transaction ID or hash for instant verification |
| Activation | "Payments are verified automatically and activate your plan immediately" |
| Regional note | "Availability varies by region" |
| Tax and fees | "Taxes, bank charges, and currency conversion fees are yours to pay" |
The tax line matters for budgeting: the listed prices are the plan prices, and the Terms say taxes and conversion fees are the buyer's. Expect the checkout total to be the figure that counts. Also note that unpaid orders "expire automatically after the window shown at checkout" and cannot be revived: you start a new one.
Two windows, two things: 7 days for a first subscription, 14 days for unused balance.
The wording differs by page, so here it is side by side. The pricing page says: "a 7-day money-back guarantee on all paid plans". The Terms (section 6) say: "Unused prepaid balance can be refunded within 14 days of the top-up, less any transaction fees, provided the account is in good standing." The Terms do not repeat the 7-day guarantee; the separate Refund Policy page does.
| Item | Window | Condition | Source |
|---|---|---|---|
| Paid plan (first subscription) | 7 calendar days | "applies only to your first subscription"; not renewals or plan changes unless a platform technical failure prevented use | Refund Policy, section 1 |
| Unused prepaid balance | 14 days from top-up | Less transaction fees; account in good standing | Terms section 6; Refund Policy section 2 |
| Consumed tokens | None | "Tokens already consumed are not refundable" | Terms section 6 |
| Billing error by Radox | Contact within 30 days | Refunded in full when confirmed | Refund Policy, section 3 |
| Crypto payments | n/a | "non-refundable in fiat currency", except verified platform-error credits | Refund Policy, section 4 |
Match the plan to the workload, not the label.
| Situation | Why it fits | Likely plan |
|---|---|---|
| Developer evaluating the endpoint | No card, all chat models | Free |
| Hobby project or prototype | Low fixed price, 100-200M pool | Starter or Basic |
| Small production app | 350M pool and 600 req/min | Pro |
| Several services in one account | 500M pool and 1,200 req/min | Scale |
| Continuous agent workloads | No token cap, 3,000 req/min | Unlimited |
Real constraints, taken from Radox's own pages.
Unused monthly tokens "do not carry over", and paid plans are charged in advance.
After the pool, you pay each model's own rate, which can be several times the plan's per-1M figure.
7 days for a first subscription, 14 days for unused balance, and nothing for consumed tokens.
Crypto payments are "non-refundable in fiat currency" except verified platform-error credits.
Even Unlimited is capped at 3,000 requests and 5M tokens per minute.
The Terms say models may be "added, renamed, repriced, deprecated, or withdrawn", and rates "can change" with advance announcement.
Billing behavior beyond the vendor's pages is unverified by third parties.
Match the need before comparing prices.
| If you mostly need... | Look at |
|---|---|
| A direct contract and terms with the model maker | The model maker's own API pricing page (Anthropic, OpenAI, Google) |
| Another hosted multi-model gateway | OpenRouter pricing |
| A gateway you host and control yourself | LiteLLM proxy (open source) |
| A first look at Claude before choosing an API | Our Claude review |
How this guide was made.
Radox AI calls to action on this page use JAVIS's referral link; if you sign up through them, JAVIS may receive a referral reward at no extra cost to you.
This is an editorial guide built from Radox AI's official pricing page, Terms of Service and Refund Policy, dated Oct 15, 2026. JAVIS has not run a hands-on test of the service. This article focuses on chat/text.
Start on the free plan; no card required per the pricing page.
Create a free accountRead the related guides below first.
Choose the next articleIf you are weighing models and tools before picking a gateway.
Questions people ask before choosing a plan
How much does Radox AI cost?
Free ($0, 10M tokens), Starter $4.50, Basic $8.50, Pro $12.50, Scale $15.50 and Unlimited $50 per month, per the pricing page on Oct 15, 2026.
What does the free plan include?
10M tokens per month, 60 requests per minute, all chat models, and community support. No card is required.
What happens when I use up my tokens?
Requests fall back to your account balance and are billed per token at each model's rate. If the balance is empty, requests return HTTP 402 until you top up or upgrade.
Do unused tokens carry over?
No. The Terms say the allowance "resets each billing period and unused tokens do not carry over". Prepaid balance, by contrast, does not expire while your account is open.
Is yearly billing cheaper?
Yes. Each yearly price equals ten months of the monthly price: Starter $45 instead of $54, Pro $125 instead of $150, Unlimited $500 instead of $600.
Is "unlimited" really unlimited?
There is no monthly token cap, but per-minute limits apply: 3,000 requests and 5M tokens per minute, with custom limits on request.
Can I change plans later?
Yes. Upgrades take effect immediately and refresh your allowance; downgrades apply from the next billing period.
Which payment methods are accepted?
Paddle (cards, Apple Pay, Google Pay, iDEAL), Binance Pay, and cryptocurrency (USDT/BTC). Availability varies by region.
Can I get a refund?
A first paid subscription can be refunded within 7 calendar days; unused prepaid balance within 14 days of the top-up. Tokens already consumed are not refundable.
Where this came from, and the date it applies to.
Plans, prices, limits, yearly pricing, payment methods, billing FAQ: radoxai.com/pricing, Oct 15, 2026.
Billing layers, payment, refunds, rate changes: Radox Terms of Service, effective 23 August 2026.
Refund windows and conditions: radoxai.com/refund, Oct 15, 2026.
Model rates used in worked examples: radoxai.com homepage model list, Oct 15, 2026.
User sentiment: no independent customer reviews were found on Oct 15, 2026; vendor statements are labeled as such.
Arithmetic: per-1M figures, yearly savings and break-even values are JAVIS calculations from the listed prices.
Images: logo hosted on radoxai.com. Official video: none identified for this product.
