Pay as you go. From $0.20 per agent hour.
One plan, one wallet. No seats, no subscriptions, no minimums. Pay for the time Antinomy agents spend working on your codebase, plus the model tokens they use, or bring your own keys and pay for agent hours alone.
Startup
For teams that would rather ship than negotiate a contract.
Model tokens billed separately from the same wallet. A month with no usage costs $0.
- Unlimited team members, projects and issues. No per-seat fees
- Every Antinomy agent, as many in parallel as you need
- No charge while an agent waits on you
- Context engine with no repository size limit
- Unlimited webhooks and scheduled runs
- Web search included at no extra cost
- Any LLM provider and model, or bring your own keys
- Cloud machines from 1x to 8x, or bring your own
No subscription, no minimum. Pay only for what your agents use.
Estimate your bill
Pick a cloud machine size and drag the slider to the number of agent hours your team expects to use in a month.
Running agents on your own machines? They are always billed at the 1x rate, whatever their size, so pick 1x.
- Rate (1x machine)
- $0.20 / hour
- Usage
- 40 hours
$0.20 multiplied by the hours your agents spend working. Model tokens are billed separately from the same wallet, or by your provider if you bring your own key.
Frequently asked questions
An agent hour is one hour of an Antinomy agent actively working on a task you gave it. Your runtime usage for the month is the total time your agents spent working, added up across every task and every agent, multiplied by the hourly rate of the machine they ran on. On a 1x machine that is $0.20, so a task that keeps an agent busy for 30 minutes costs half an agent hour, or $0.10.
No. Agent hours only count the time an agent is actively working. If an agent pauses to ask you a question or wait for your approval, the clock stops, and that waiting time is not charged. It starts again once you respond and the agent gets back to work.
Two things, billed separately: agent runtime and model tokens. Agent runtime is the time your agents spend working, charged per agent hour. Model tokens are the prompts and completions your agents send to the LLM provider, charged by usage. Both are deducted from the same wallet.
Your workspace has a single shared wallet. You top it up with credits, and every charge, whether agent hours or model tokens, is deducted from that one balance. There is nothing to reconcile across team members or agents: one wallet, one balance, one place to look.
Yes. Add your own key for any supported LLM provider and Antinomy will route model calls through it. In that case your provider bills you for the tokens directly, and the only thing deducted from your wallet is agent hours.
Agents on Antinomy cloud run on a machine size you choose: 1x, 2x, 4x or 8x. The hourly rate scales with the size, so an agent hour on a 4x machine costs four times the 1x rate. Use the estimator above to see what your expected hours cost at each size.
Yes. Bring your own machine and Antinomy will run agents on it instead of on our cloud. Agent hours on your own hardware are always billed at the 1x rate, no matter how many cores the machine actually has, and model tokens are charged the same way as everywhere else, or not at all if you bring your own key.
Each running agent is counted separately and the time is added together. Three agents working in parallel for 20 minutes each add up to one agent hour, the same as a single agent working for an hour. You pay for the work done, not for how many agents did it.
No. Web search is included at no extra cost. There is no per-search charge and nothing is deducted from your wallet for running a search. You pay only the usual agent hours and model tokens for the time and context the search uses.
No. There are no seats, no subscriptions and no minimums. Add your whole team and pay only for the agent hours and tokens you actually use. A month with zero usage costs $0.
Ready to Get Started?
Join thousands of developers already using Antinomy. Start free today.