Terminaili
Menu
Pricing

Pick a monthly bundle. Stay in control of model spend.

Bundles are calibrated so every $10 covers ~$5 of model usage at retail across every class. Cheap models stretch your bundle; premium models burn it faster. No model class is the special one - a Pro plan stays a Pro plan.

Get startedView docs
Entry

Starter

$10/mo1 bundle

Light usage, evaluation, and small hobby projects.

  • 2M input bundle tokens
  • 500k output bundle tokens
  • Credit top-ups available
You actually get (in / out, per month)
Router
10M / 2.5M
nano
33M / 8.3M
gpt-5.5
400k / 100k
5.5-pro
67k / 17k
Start Starter
Regular

Builder

$20/mo3 bundles

First real paid tier — comfortable for regular weekly work.

  • 6M input bundle tokens
  • 1.5M output bundle tokens
  • Credit top-ups available
You actually get (in / out, per month)
Router
30M / 7.5M
nano
100M / 25M
gpt-5.5
1.2M / 300k
5.5-pro
200k / 50k
Start Builder
Daily

Pro

$30/mo6 bundles

Where daily individual coding-agent use starts feeling credible.

  • 12M input bundle tokens
  • 3M output bundle tokens
  • Credit top-ups available
You actually get (in / out, per month)
Router
60M / 15M
nano
200M / 50M
gpt-5.5
2.4M / 600k
5.5-pro
400k / 100k
Start Pro
Heavy

Power

$40/mo10 bundles

Heavy individuals and small teams. The don't-think-about-it plan.

  • 20M input bundle tokens
  • 5M output bundle tokens
  • Credit top-ups available
You actually get (in / out, per month)
Router
100M / 25M
nano
333M / 83M
gpt-5.5
4M / 1M
5.5-pro
667k / 167k
Start Power
One bundle1 bundle = 2M input + 500k output bundle tokens. Bundle tokens are a normalized unit - each model class draws them at its own burn rate.
Model burn ratesBurn rates are calibrated so every $10 bundle covers ~$5 of model usage at retail across classes. Router 0.2x, gpt-5-nano 0.06x, gpt-5.5 5x, gpt-5.5-pro 30x. The 'You actually get' box on each plan card shows the exact monthly tokens per model.
OverageOut of bundle tokens? Buy another bundle at $10 (2M input + 500k output) or let credits cover ad-hoc spillover.
Credits still workPrepaid credits ($0.04 each) remain available for top-ups, ad-hoc model calls, and add-ons - including overage on bundles.
No unlimited trapBundles cap at the level you pick. Overage and credit top-ups are opt-in, never auto-billed without limit.
Telemetry unitsEvery organization gets 1,000,000 App Telemetry units per month. Beyond that, units draw from bundle tokens or credits.
Hosted Apps computePreview and idle app compute is 1.25 credits/hour for Quick, 2.5 for Standard, and 6.25 for Large while running, building, or waking.
Model costs stay visibleThe dashboard keeps showing per-model input, output, and burn-rate usage so spend is never abstracted away.