Pricing
Pay for capability, not for patience.
There is no free plan and no throttled tier. Every plan is the full product with a generous allowance, and the only thing that changes as you go up is how much of it you get. Pay yearly and every tier is 20 percent less.
- Generous Code/CLI usage, sized to this plan
- Chat shares the Code/CLI 5-hour and weekly pool
- One rate card for every API key
- Echo and Horizon, 1M context, from the same pool
- Overage from prepaid credits, only if you say so
- Generous Code/CLI usage, sized to this plan
- Chat shares the Code/CLI 5-hour and weekly pool
- One rate card for every API key
- Echo and Horizon, 1M context, from the same pool
- Overage from prepaid credits, only if you say so
- Generous Code/CLI usage, sized to this plan
- Chat shares the Code/CLI 5-hour and weekly pool
- One rate card for every API key
- Echo and Horizon, 1M context, from the same pool
- Overage from prepaid credits, only if you say so
- Generous Code/CLI usage, sized to this plan
- Chat shares the Code/CLI 5-hour and weekly pool
- One rate card for every API key
- Echo and Horizon, 1M context, from the same pool
- Overage from prepaid credits, only if you say so
- Generous Code/CLI usage, sized to this plan
- Chat shares the Code/CLI 5-hour and weekly pool
- One rate card for every API key
- Echo and Horizon, 1M context, from the same pool
- Overage from prepaid credits, only if you say so
One bucket. Switch freely.
Editor and terminal share one generous allowance, sized to the plan you pay for. Start a refactor in the IDE, finish it over SSH, and the meter does not care.
Thinking is not billed as shipping.
Chat has its own generous allowance, independent of Code/CLI and also sized to the plan. A long architecture conversation never costs you a fix.
Prepaid. Opt-in. Yours.
When allowance runs out you choose: hard-stop, or spend credits you already bought. API keys always use credits, so production is a number you set.