Pricing
Pay for what you use.
Model usage is metered per token against published per-model rates. The access period below covers the console, keys, logs and routing — never the tokens themselves.
Access periods
One flat fee for the platform, tokens billed separately.
No automatic renewals. Pick the period that matches how long you are building.
| Access period | Price | Effective per day | |
|---|---|---|---|
| DailyFocused access for short sessions and evaluation. | $3per 24 hours | $3.00 | Get access |
| WeeklyFlexible access for active builds and sprints. | $8per 7 days | $1.14 | Get access |
| MonthlyConsistent access for regular development. | $22per 30 days | $0.73 | Get access |
| YearlyLong-term access at the lowest daily cost. | $199per 365 days | $0.55 | Get access |
Access periods exclude token consumption, which is billed at the per-model rates listed below.
What a month costs
Estimate against a real workload.
Assumes a 62/38 prompt-to-completion token split, which is typical for chat and agent traffic.
Per-model token rates
USD per 1,000,000 tokens. Input and output are metered separately.
| Model | Context | Input / 1M | Output / 1M | Access | Monthly estimate | |
|---|---|---|---|---|---|---|
| Claude Sonnet 5claude-sonnet-5 | 1M | $3.00 | $15.00 | All plans | $378.00 | Details |
| GLM 5.3 Flashglm-5.3-flash | 256K | $0.25 | $1.10 | All plans | $28.65 | Details |
| GLM 5.2glm-5.2 | 256K | $0.60 | $2.20 | All plans | $60.40 | Details |
| DeepSeek V4 Prodeepseek-v4-pro | 256K | $0.90 | $3.40 | Model add-on | $92.50 | Details |
| DeepSeek V4.1 Flashdeepseek-v4.1-flash | 256K | $0.20 | $0.80 | All plans | $21.40 | Details |
| GPT 5.6 Lunagpt-5.6-luna | 1M | $2.50 | $10.00 | Premium wallet | $267.50 | Details |
Estimates are illustrative. Actual cost depends on the prompt-to-completion mix of your traffic.
Other ways to buy
Three options beside the access period.
Token packs
Prepaid tokens that do not expire, usable across the core catalog. Useful when usage is bursty rather than continuous.
Model add-on
Unlocks the advanced reasoning and coding models on any active access period, billed per token like the rest of the catalog.
Premium wallet
A separate balance for the highest-capability multimodal models, so a premium experiment cannot exhaust your main allowance.
Included with every plan
No feature is held back behind a tier.
Every modelThe full catalog at the same per-token rate on any access period.
Console accessKeys, logs, analytics, usage and routing configuration.
Both API formatsOpenAI chat completions and Anthropic messages.
Fallback routingRetry and fallback chains on every request.
Request-level logsFull payloads, headers and timing per request.
Usage meteringTokens and spend attributed per model and per key.
Not sure which fits?
Compare the three purchasing models side by side.
Access period for continuous use, token packs for bursty use, and the premium wallet for expensive models.
Billing rules
The short version.
- RenewalsManual. Nothing charges automatically.Included
- RefundsReviewed when access was not delivered.Included
- Unused tokensToken packs do not expire.Included
- Failed requestsNot billed when they never reach a provider.Included
Ready to send your first request?
Create a key and point an existing client at one base URL.