Free Tier Watch

What will this AI feature cost?

Describe one call and how often it happens. This works out the monthly bill at current list prices, and tells you whether it fits inside the free quotas — which is the part that decides whether a side project can exist at $0.

Loading…
10,000 is roughly 330 a day.
A page of context is roughly 750 tokens.
A short paragraph is roughly 100 tokens.
All models are still compared below.

$0.00 / month

Monthly cost by model

At the workload above, cheapest first. Hover a bar for the per-call figure.

    List prices per million tokens, verified —. Cache-hit pricing is excluded — where a provider offers it, repeated context can cost an order of magnitude less than shown.
    Show as a table
    Monthly cost by model at the workload above
    ModelVendor In $/M Out $/M Per call Per month

    Does it fit in a free quota?

    Checked against your calls-per-month figure above. Free quotas are the reason a hobby project can run at zero marginal cost — and the ceiling you hit the moment it gets popular.

    Prices move. Get told when they do.

    A short weekly email covering model price changes, free-quota changes, and platform terms updates.

    One email a week. Unsubscribe in one click.

    Read this before trusting the number. These are list prices for a single call shape, so they are a floor, not a forecast — retries, failed parses, system prompts, and long conversation histories all push real spend up. Some prices here are from a secondary aggregator rather than the vendor's own page; those rows are marked in the table. Google no longer publishes its free-tier rate limits at all, so the Gemini free quota below cannot be verified from documentation — you have to read your own dashboard.