generation_spend_cap_usd of 0, 120, 450, null or 600: which applies?
On a Format run 120 and 450 are used as given, null is the $500 platform maximum, and 0 or 600 returns a 400. Omitted means the Format's cap or $400.
On a Format run, generation_spend_cap_usd: 120 caps the run at $120, 450 caps it at $450 even if the Format's own cap is lower, null lifts it to the $500 platform maximum, and 0 or 600 is a 400. If you omit the field, the run inherits the Format's cap, which is $400 when the Format never set one.
The table
This is the resolution rule from the Format call guide.
| You send | The run's cap |
|---|---|
| Nothing | The Format's cap, or $400 if it never set one |
| 120 | $120 |
| 450 | $450, even if the Format's cap is lower |
| null | $500, the platform maximum |
| 0 | 400 error |
| 600 | 400 error |
What a cap does and does not do
The Format errors page describes two gates. The wallet gate runs at create and returns 402 insufficient_credits if the workspace cannot fund the run. The spend cap applies during the run; a run that hits it ends as failed, and usage shows how near to the cap the spend got.
The cap is not a bill. The receipt's usage.billable_amount_usd_micros counts reserved plus captured generation amounts and excludes the agent's own LLM turn, so it is not the total cost. Use GET /v1/usage and GET /v1/balance as the billing records.
Picking a number
Production long-form runs usually use caps near $120, and a single-scene retry needs a few dollars. Size a cap from the rate card: a run that makes six 15 second plus avatar clips needs 6 x 15 x $0.245 = $22.05 before images, music and the timeline, so a $30 cap would be tight and $120 would be generous.
Because the API accepts a request cap above the Format cap and does not clamp it, treat the request field as a deliberate override, not a default.
Two examples
A Format has a cap of $200. A request that sends nothing runs under $200. A request that sends 450 runs under $450, because the API accepts a number above the Format's cap and does not clamp it. A request that sends null runs under $500. A request that sends 600 is rejected with 400, and so is 0.
A Format that never set a cap reports the platform default of $400 in generation_spend_cap_usd_micros. The receipt shows the effective cap as usage.generation_spend_cap_usd_micros on every run, which is the value to log.
- Cap in USD micros: $120 is 120,000,000.
- Do not confuse the cap with the wallet check at create.
- A run that reaches the cap ends as
failed.
Sources
Related posts
More in Pricing
- GLM-5.3 off-peak 50% points vs a Sume dollar spend cap
Z.ai meters GLM-5.3 in points with a 50% off-peak rate. Why a Sume dollar cap is a different meter, and how to schedule agent work.
- GPT Image 2.5 on fal runs $0.00402 to $0.40026 per image: on Sume
fal's ChatGPT Image 2.5 Flare page lists $0.00402 for 1024x768 low and $0.40026 for 3840x2160 max. At Sume's 1.25 multiplier that is about $0.005 to $0.50.
- Does gpt-image-2.5 cost more at 4K? Prices from 1024 to 3840x2160
gpt-image-2.5 price by pixel size on Sume: 1024x1024, 1536x1024, 2048x2048 and 3840x2160 at medium and high, plus the custom-size rules from the docs.
- GPT Image 2.5: leave quality blank and you pay 7 cents, not 1
On Sume, GPT Image 2.5 defaults to high quality when you omit the field. A 1024-class image is 7 cents at high, 2 at medium and 1 at low. 100 images: $7 vs $1.
Written by Sume