Billing & plans

AlayaCode is subscription-based — fixed monthly quota, no per-call surprise charges. This page covers four things: picking a tier, how tokens are counted, when rate limits trigger, and what happens when you run out.

Tiers

Tier	Monthly	Token quota / mo	RPM	TPM	Models	For
Free	¥0	1 M	60	30 K	minimax-m2.1, glm-5 (throttled)	Trial, personal hacks
Pro	¥X.XX	30 M	300	150 K	+ minimax-m2.5, glm-5.1	Daily-driver for individual devs
Team	¥X.XX	200 M	600	500 K	+ deepseek-v4-pro (1M ctx)	5-20 person teams
Enterprise	Contact sales	Custom	Custom	Custom	All + private custom models	Large teams / high-compliance

Live prices are in the console; promotional credits, quarterly commitment rebates, etc. follow the marketing calendar.

¥X.XX is a placeholder. Pricing is set by ops, not docs — we don't hard-code numbers here. The console (AlayaCode → Plan) shows the live rate.

How tokens are counted

Input and output tokens cost the same, scaled by the model's pricing factor:

quota_used = (input_tokens + output_tokens) × model_factor

Sample factors:

Model	factor
`minimax-m2.5`	1.0
`minimax-m2.1`	0.6
`glm-5`	1.0
`glm-5.1`	1.2
`deepseek-v4-pro`	2.0 (long-context model)

Example: on Pro (30 M tokens / mo), a glm-5.1 call with 500 input + 800 output tokens costs (500 + 800) × 1.2 = 1560 quota tokens.

Rate limit triggers

Two dimensions, whichever hits first:

RPM (requests per minute), 60 s rolling window
TPM (tokens per minute), 60 s rolling window, summing input + output

A trigger returns 429 rate_limit_exceeded with a Retry-After header on the next request. Throttling does not consume quota.

Quota exhausted ≠ service down

When your monthly token quota is spent, we don't error by default — we drop you into "throttled mode":

Mode	Behavior
Normal	RPM / TPM at tier ceiling
Throttled	RPM cut to 60, TPM cut to 30 K (same as Free tier)

Throttled-mode requests come from a separate "throttle pool" (refills 100 K / day, same across tiers). If even that drains, you get 429 quota_exhausted.

Want to avoid throttling? Two options:

Buy a one-shot top-up in the console — quoted in tokens, billed at the tier's factor
Upgrade tier — proration applies for the rest of the month

Usage monitoring

AlayaCode → Usage in the console shows:

Spent / total quota this month
Consumption split by model
Daily call volume curve
Headroom before throttling
Hourly rate-limit-trigger count

Email alerts can fire at 80% / 95% quota.

Invoices & enterprise payment

Subscription billing draws from the main account balance; top-up methods and invoice requests live on the billing reference.

Cancellation

AlayaCode → Plan → Downgrade / Cancel. The current cycle's fee is not refunded — change takes effect at the next billing cycle.