Pricing

Server Fee

Charged per hour for the machine your agent runs on. Starter is the default at launch; you can upgrade to a tier with more cores when you need it.

Tier

Spec

Price

Starter

2 vCPU · 4 GB

0.05 USD/hr

Premium

4 vCPU · 16 GB

0.19 USD/hr

Max

8 vCPU · 32 GB

0.37 USD/hr

Billed only while the instance is running. Stopping the instance stops the server fee.

Data Fee (Blave API)

Charged when your agent pulls market data through the Blave data API. BlaveClaw fetches quotes and analytics from these endpoints as it runs.

0.10 USD/hr — charged once for any hour in which the Blave data API is called.

Charged at most once per hour — multiple calls in the same hour are not double-charged.

API plan subscribers are exempt — the API plan already unlocks Data API access, so data is included in the subscription.

LLM Fee

Charged per token by the model your agent uses. Sonnet is the default. Rates below are in USD per 1M tokens.

Model	Input	Cache Write	Cache Read	Output
Claude
Haiku	1.33	1.67	0.13	6.67
Sonnet	4.00	5.00	0.40	20.00
Opus	6.67	8.33	0.67	33.33
Fable	13.33	16.67	1.33	66.67
DeepSeek
DeepSeek V4 Flash	0.18	0.18	0.004	0.37
DeepSeek V4 Pro default	0.55	0.55	0.005	1.10

Per 1M tokens, in USD.

Web search is available only on Claude (Anthropic) models — DeepSeek does not bill for it. Each search costs 13.33 USD / 1000 searches.

DeepSeek Peak Hours DeepSeek rates double during Beijing peak hours (09:00–12:00 and 14:00–18:00, UTC+8).

To switch models, just tell BlaveClaw directly — e.g. "Use Claude Sonnet from now on."

Want other data? Come find us in our Discord community and let us know what you're looking for!