Billing & Credits

How Credits Work

Token101 uses a prepaid credit model. The conversion rate is simple: 1 credit = $0.01 USD. For request execution, credits are temporarily held before a request is processed and adjusted after actual usage is calculated. If the final charge is lower than the amount held, the unused portion is returned automatically.

Credit expiry depends on how they were granted: subscription and weekly-quota credits expire at the end of their period and do not roll over, while purchased top-up credits do not expire by default. The register gift expires 30 days after it is granted.

---

Subscription Plans

Token101 offers four subscription tiers, billed monthly — compare them side by side on the pricing page:

Free Plan Details

• Register gift: One-time 50 credits after email verification (the free plan has no recurring monthly credits)
• Gift expiry: The register gift expires 30 days after it is granted, so claim it and start building early; the exact amount and expiry window follow the current promotion settings and may change over time
• Model access: All active models — gift credits work everywhere; the free tier's spend cap keeps usage bounded

Spend Caps

Each plan has weekly and monthly credit spending caps that limit how many credits can be consumed in a window, regardless of your total balance:

If you hit a cap, API requests return a rate-limit error until the cap window resets (anchored to your billing period).

---

HotPricing Plans

For users who need higher volume, HotPricing offers PAYG packages and weekly-reset subscriptions with significantly larger credit allocations and spending caps.

HotPricing PAYG (One-Time Top-Up)

HotPricing PAYG credits are granted once and do not expire — they remain available until fully consumed. Each package also unlocks a higher spend cap, allowing more API calls per window.

HotPricing Subscriptions (Weekly Reset)

HotPricing subscriptions grant credits on a weekly reset cadence — you receive a fresh allocation every week, and unused credits expire after 7 days. This model is designed for heavy, consistent usage where weekly budgeting matters more than rolling balance.

VIP Tiers

Cumulative spending unlocks VIP service tiers with priority support:

VIP is currently a service benefit tier (billing discounts are reserved for a future phase and are not applied at checkout today).

---

Standard PAYG Top-Up

If you don't need HotPricing volumes, standard PAYG top-ups are available in USD:

First-time top-ups receive a 10% bonus (bonus credits expire after 30 days). Subsequent top-ups receive the base credit amount only.

API Activation Pack

Free users who want to create API keys must first complete API activation. A special ¥4 activation pack grants 10 credits — enough to start using the API. This pack is only available from the API Keys settings page and is not listed on the general pricing page.

---

Model Pricing

Text Models (Per-Token)

All text models are billed per 1M tokens, with separate input and output rates that vary by model. Per-token prices are shown on the Supported Models page and returned by GET /api/v1/public/models — read them there rather than relying on a static table, since prices can change as the catalog evolves.

Long-Context Pricing

Some models can apply a higher pricing tier once a request crosses a large token threshold (for example, 200,000 tokens). Among the models currently available, none charge a higher rate for long context — the large-context models use flat per-token pricing regardless of prompt length. Always confirm the current rate on the Supported Models page.

Reasoning Token Discounts

Models that produce reasoning/thinking tokens (typically Codex and advanced OpenAI models) receive an automatic discount on reasoning output:

This means reasoning tokens are billed at a reduced output rate. For example, on gpt-5.3-codex with a $10/1M output rate, reasoning tokens cost $5/1M instead.

Prompt Caching Costs

Models that support prompt caching automatically reuse repeated request content:

• Cache hits (reads): cached portions are charged at only 10–20% of the input price (an 80–90% saving).
• Cache writes: content written into the cache for the first time is charged at 1.25× the input price (a 25% premium). This is why a line on your bill can show a unit price slightly above the standard input rate — it corresponds to a cache write, and later hits recover that cost.
• Scope: only models that support prompt caching (such as Claude and the GPT 5.x series) carry these adjustments; models without caching support are billed at the standard input price with no cache premium or discount.
• Savings: the amount caching has saved you is shown on the Usage page.

Usage boundary: when you use the Anthropic cache_control field, omit ttl to use the default 5-minute cache; setting ttl: "1h" explicitly is rejected (HTTP 400), so leave the ttl field out.

Image Generation (Per-Image)

Image generation uses per-image pricing instead of per-token billing. The price varies by quality and size:

Within the same quality tier, non-square sizes cost slightly more than 1024×1024. Confirm the current price from the public models API; see Images for the request parameters.

---

Viewing Your Usage

Use Dashboard > Usage to view your plan, balance, and quota state. The /api/v1/me/* endpoints below power the Dashboard and require a logged-in browser session — they cannot be called with an API key. For programmatic model discovery, use GET /api/v1/models with your API key. Usage and balance figures are currently available in the Dashboard only.

---

Billing API Endpoints