Runars uses a simple, transparent credit-based billing model. Pay only for what you use, with no contracts or recurring fees. All prices are in Canadian dollars (CAD); check the live catalog for current rates.

Model pricing

Each model in the catalog has its own per-token rates. The default glm-5.2 model:

Metric	Price (CAD)
Input tokens (per 1M)	$1.00
Output tokens (per 1M)	$4.00

Prompt caching

If you're using prompt caching (available via the Anthropic SDK and compatible clients), cache tokens are billed at a different rate:

Cache Type	Rate
Cache write (5-minute TTL)	1.25× input token price
Cache write (1-hour TTL)	2× input token price
Cache read	0.1× input token price

This mirrors Anthropic's own prompt-caching economics, so switching to Runars doesn't change how you think about caching costs.

Web search

Web search (available on /v1/messages and /v1/responses) is billed per executed search, in addition to normal token costs:

Metric	Price (CAD)
Web search (per 1,000 searches)	$10.00 USD equiv.

This is the standard rate Anthropic and OpenAI charge for their hosted web search tools — Runars charges it with no markup and executes the searches gateway-side. Errored searches (provider outage, max_uses cap) are not billed. Search-result content fed to the model bills as normal input tokens.

Billing model

Credit-based: You prepay for credits and they're deducted as you make requests.
Transparent: Each request shows exactly how many tokens were used and how much it cost.
No overage charges: Your account stops accepting requests when your balance reaches zero (402 payment_required error).
Organization-scoped: Credits belong to an organization, not a user. Everyone in your org shares the same balance.

Adding credits

To add credits:

Go to Dashboard → Payments
Click Add Funds
Choose an amount between $5–$500
Complete the Helcim-powered checkout
Credits are applied immediately

Receipts are sent to your registered email via Resend.

Rate limits

Rate limits are tied to your organization and scale with your lifetime spending:

Tier	Lifetime Spend	Requests/Minute	Input Tokens/Minute	Output Tokens/Minute
Start	$0	1,000	2,000,000	400,000
Build	$100+	5,000	5,000,000	1,000,000
Scale	$1,000+	10,000	10,000,000	2,000,000
Enterprise	$5,000+	20,000	10,000,000	2,000,000

Your current tier is displayed in Dashboard → Rate Limits, along with your progress toward the next tier.

Rate limit errors

If you exceed a rate limit, the API returns:

429 Too Many Requests (Anthropic-compatible endpoints /v1/messages, /v1/messages/count_tokens)
429 insufficient_quota (OpenAI-compatible endpoints /v1/chat/completions, /v1/responses)

Both include a retry-after header indicating how long to wait before retrying.

What counts against rate limits

Requests: Each API call (even non-streaming requests).
Input tokens: All tokens in your request's messages, including system prompts and previous conversation history.
Output tokens: Only tokens actually returned by the model (not your max_tokens limit).

The count_tokens endpoint (POST /v1/messages/count_tokens) does not count against rate limits—use it freely to pre-estimate token costs.

Tracking usage

Visit Dashboard → Usage to see:

Current credit balance
Usage this month (tokens and cost)
Lifetime spending
Billing history

Check Dashboard → Logs to see request-level details.

Pricing