Billing & Charges — Common Questions

Was I charged for a failed generation?

It depends on when the failure happened:

Failure point	Charged?	Why
Before submission — validation error, missing media, API key invalid	No	The request never reached fal.ai
During queue / startup — model timed out, server error	Usually no	fal.ai typically does not charge for generations that produce no output
During generation — content policy flag, model crash	Usually no	Most models only bill on successful output
After generation — download failed, import failed	Yes	The generation completed on fal.ai’s side. The Billing tab marks these with a red failed-run badge and counts them toward your total. Clicking Generate submits a new request, which fal.ai may charge for — nothing in the plugin re-fetches the earlier result

When in doubt, check your fal.ai billing dashboard for fal.ai’s own aggregate billing records. modelBridge shows “Were you charged?” guidance on every error in the Error Reference.

Why does the cost in modelBridge differ from my fal.ai dashboard?

Three common reasons:

1. Estimate vs. actual. The cost badge before generation is an estimate based on the best available pricing data. fal.ai’s final charge may differ — especially for models where pricing depends on output complexity or internal provider logic, not just input parameters. After generation, the Billing tab shows the metered cost when fal.ai reports usage modelBridge can price; otherwise it shows modelBridge’s own calculation, labelled Calculated.

2. “From” badge. When the cost badge says “From $X.XX,” this is the minimum starting price. Audio, higher resolution, or longer duration can increase the final charge above the displayed minimum. See what “From” means.

3. Currency conversion. If you display costs in a non-USD currency, the converted amount may differ from the USD amount on fal.ai’s dashboard due to exchange rate timing. modelBridge uses daily rates from the European Central Bank.

fal.ai’s billing is always the source of truth for what you’re charged. The Billing tab’s Metered figure is the units fal.ai reported, priced at fal.ai’s own rate for them — modelBridge’s multiplication, not read from your fal.ai invoice.

How do I set a spending limit on fal.ai?

fal.ai supports spending limits directly on their platform:

1. Go to fal.ai/dashboard/billing
2. Look for the spending limit or budget cap settings
3. Set a monthly or total limit

When your balance or limit is reached, generations will fail with a “Balance empty” error. modelBridge will show a clear message with a direct link to add funds.

modelBridge does not enforce spending limits itself — this is controlled entirely by your fal.ai account.

Where do I see my total spending?

Inside modelBridge: Open the Billing tab. The running total at the top shows your cumulative AI spending across all generations. Filter by project to see per-client costs. See Billing Tab for the full walkthrough.

On fal.ai: Your fal.ai billing dashboard shows the authoritative transaction history and current balance.

Export: The Billing tab supports CSV export and HTML report generation for accounting and client invoicing. See Cost Tracking & Reports.

What happens to my fal.ai credits if I cancel modelBridge?

Nothing. Your fal.ai account is completely independent from your modelBridge subscription. Cancelling modelBridge does not affect your fal.ai balance, API keys, or account in any way. You can continue using fal.ai through their website, API, or other tools.

See Subscription & Account for full details on what happens when you cancel.

Can I get a refund for a generation I am unhappy with?

Refunds for AI generation costs go through fal.ai, not modelBridge. modelBridge does not handle or process the payment for generations — fal.ai charges your account directly.

If a generation failed due to a platform error (not your input), fal.ai typically does not charge. If you were charged for a genuinely broken result, contact fal.ai support.

For modelBridge subscription refunds, see the Refund Policy.

Why was I charged more than the estimate showed?

The most common reasons:

• Audio was enabled. Many video models charge significantly more with audio generation. Check whether audio was toggled on.
• Resolution surcharge. Higher resolution tiers (1080p, 4K) cost more than the minimum. The “From” badge reflects the lowest tier only.
• Duration scaling. Longer durations multiply the per-second rate. A 10-second clip costs roughly twice as much as a 5-second clip.
• Provider pricing changed. fal.ai can update model pricing at any time. modelBridge re-reads a model’s published rate the next time you open that model, so there can be a window between a price change on fal.ai and modelBridge reflecting it — and a model you don’t open isn’t re-read at all.

To minimize surprises: prefer models with the Estimated badge (most accurate pre-generation tier), check the cost breakdown before generating, and run a short test generation at lower resolution before committing to expensive settings.

Why can’t I see my fal.ai balance inside modelBridge?

modelBridge does not display your fal.ai account balance. Reading it would require an admin-scope key, and modelBridge asks only for an API-scope key — the narrowest one that can run a generation. Your balance and transaction history stay where your credentials are already authenticated.

If your balance runs out mid-session, modelBridge will detect this when a generation fails and show a clear message with a direct link to top up your fal.ai credits. Once topped up, return to modelBridge and try again.

To check your balance: fal.ai/dashboard/billing