Simple, transparent pricing

One Gateway. Pay only for what you use — at exact provider rates, with no markup.

Working as a team? Compare Free, Teams & Enterprise plans.

Free

$0/forever

Perfect for getting started.

  • Access to default free models
  • Bring your own keys (BYOK)
  • Basic usage analytics
  • No credit card required
Get started free
Most popular

Pay-as-you-go

Market rates

For serious developers and production apps.

  • 500+ models across 60+ providers
  • Exact provider rates with no markup
  • No markup on BYOK
  • Advanced analytics & load balancing
  • Top up credits only when you need to
Start building

Pass

From $19/month

Maximize your AI output per dollar.

  • Monthly AI credit subscription
  • Up to 50% bonus credits
  • Credits work across all routed inference
  • Auto Model stretches credits further
  • Cancel anytime
Get Pass
MegaBrain Plan

One subscription, one agent, rolling limits

A flat monthly price with 5-hour, weekly and monthly usage windows instead of a metered bill. Hit a window and the agent keeps going on Brain Core, or spills to Extra Usage if you turn that on. MComputers — the machines the agent runs on, ours or yours — start at Plus.

Pro

$20/month

The agent everywhere you already work.

  • Standard included usage
  • MComputers not included
  • Agent in the editor, CLI and chat
  • Brain Core keeps working after the limit
  • Extra Usage overflow — off by default, with a spend cap
Get Pro

Plus

$100/month

Add the machines the agent runs on.

  • ≈5× the usage of Pro
  • 1 hosted MComputer + 3 of your own
  • Bring your own machine — any Linux VPS or Mac
  • Remote control of a connected Mac
  • Webhook and cron triggers on the machine you choose
Get Plus

Max

$200/month

For running several agents at once.

  • ≈10× the usage of Pro
  • 3 hosted MComputers + 10 of your own
  • Parallel sessions across machines
  • Priority routing on the shared pool
  • Early access to new capabilities
Get Max

Included usage is shown as a share of each window, never as a dollar allowance — the plan sells a package, not tokens. Pay-as-you-go through the raw Gateway API stays a 0%-markup pass-through, and BYOK inference is never counted against your windows.

Auto Model

Choose a tier, let Auto Model route the rest

Auto Model is included on every plan. Pick the cost-control tier that fits your budget.

Frontier

Latest, most capable paid models for the hardest tasks.

Balanced

Cost-effective paid routing — capable results without frontier prices.

Free

Best available free models, traffic split across them at no cost.

Pricing FAQ

What is AI inference and why is it priced this way?+

AI inference is the cost of running your prompts through a model. We pass through exact provider rates with no markup — you only pay what the underlying provider charges, plus optional bonus credits if you subscribe to a Pass.

What is the difference between Pay-as-you-go and Pass?+

Pay-as-you-go bills usage at exact provider rates with no commitment. Pass is a monthly credit subscription that rewards you with up to 50% bonus credits for regular usage.

How does Auto Model affect cost?+

Auto Model routes each request through the model strategy that matches your selected tier, so you spend fewer credits on tasks that do not need frontier reasoning.

Do I need a subscription to start?+

No. You can start free with free models or your own keys, with no credit card required.