Veo 3.1 result exceeds ~16 MiB (undocumented inlineTooLarge)

Hi Vercel team,

We run video generation through AI Gateway's asynchronous job API (experimental_startVideo + experimental_getVideoStatus). For google/veo-3.1-generate-001 at 4K, the gateway reports a job as "completed" but returns an empty videos array whenever the rendered file is larger than roughly 16 MiB. The job's providerMetadata carries a flag that appears in no documentation:

result: { inlineTooLarge: true, expired: true, expiredAt: }

The generation finished at Google and the job carries a gateway generation id, so it was billed, but the output is unrecoverable through any API we can find.

ENVIRONMENT

ai 7.0.82, /gateway 4.0.67, Node 24.19, API key authentication
Team: API key:

REQUEST SHAPE (identical on every affected job)

model: google/veo-3.1-generate-001
prompt: { image: , text: }
aspectRatio: "9:16", resolution: "3840x2160", duration: 6, generateAudio: false
idempotency-key per job; status polled every ~6 s with experimental_getVideoStatus

AFFECTED JOBS (all "completed", videos: [], warnings: [], times in UTC)

job_4JT9B4B2SQBFN5QZC6FZN8X0QF started 2026-09-03 00:41:46 completed 00:48:01 gen_01M1JBDYXYJ0Q1QJQ859BY50X0
job_15VMJC859DS4T2DDCHAQT7TNBC started 2026-09-03 00:41:46 completed 00:48:34 gen_01M1JBDYYNZ4Z8V9JYQPTJRK9J
job_68M78EC73RVNDWQBD5XKBJ37WA started 2026-09-03 00:41:46 completed 00:48:06 gen_01M1JBDZ9FB4MC8X02M709VGZW
job_2B31C6K6ETYK4W8EF8K4EWB311 started 2026-09-03 00:50:42 completed 00:57:32
job_6QKPBBCF960RGC5NE4RGD7VERH started 2026-09-03 00:50:42 completed 00:57:09
job_09M1X5ERDDRWM7K1E1KD4E02JG started 2026-09-03 00:50:42 completed 00:57:28
job_13SWDZ26JKKK2JPYESBZ1KAH0T started 2026-09-03 00:50:42 completed 00:56:50

DELIVERED JOBS, same model and parameters, minutes apart (result: { delivery: "inline" }, videos[0].type "base64")

job_5ZW2AZV286Q1KFSPDXBKS76JT4 15.7 MB (2026-09-02 23:49 to 23:56)
job_4B8P210EMATYA76BH8CRNE6QM4 14.7 MB (00:41 to 00:48)
job_5JPFF9N4ZVE9M8YKR1WC3HKJQS 15.1 MB (00:41 to 00:48)
job_5GR1EVC3VTF6F0RZKY72BV75Q3 13.0 MB (00:50 to 00:56)

OBSERVATIONS

The largest inline payload ever returned to us is 16,123 KB. Re-running the dropped prompts dropped them again, and the delivered ones came through again, which points to a fixed size threshold rather than a transient fault.
On dropped jobs expired is already true at completion time, while expiredAt is completion + 24 h, the same window a hosted result gets. The result appears to be discarded rather than stored.
The status is "completed" with empty warnings. There is no error, no 413, and the webhook payload carries no bytes and sends the receiver back to the same status call.
The documentation mentions only the 300 KiB cap on the persisted start request and the expiresAt / expiredAt / partial / unretained keys. Neither inlineTooLarge nor a result-size limit is documented.
Providers that hand back a hosted URL are unaffected: 18 of 18 Seedance jobs in the same period delivered with result: { expiresAt }.

EXPECTED Either the result is hosted with an expiresAt, as URL providers' results are, or the job ends in an explicit error so a client can retry at a smaller size. A "completed" job with no output and no warning is indistinguishable from success.

REQUESTS a) Host inline results above the cap, or raise the cap for the async job store. b) Until then, return an error status, or at least a warning, when a result is dropped. c) Document the limit and the inlineTooLarge key. d) Credit for the seven generations billed without an output.

RAW STATUS RESPONSE for job_68M78EC73RVNDWQBD5XKBJ37WA (queried again at 2026-09-03 00:53 UTC) { "status": "completed", "videos": [], "warnings": [], "providerMetadata": { "gateway": { "asyncJob": { "jobId": "job_68M78EC73RVNDWQBD5XKBJ37WA", "status": "completed", "completedAt": 1788396479596, "result": { "expired": true, "expiredAt": 1788482879550, "inlineTooLarge": true } }, "generationId": "gen_01M1JBDZ9FB4MC8X02M709VGZW" } }, "response": { "modelId": "google/veo-3.1-generate-001", "headers": { "x-matched-path": "/v4/ai/video-model/status", "x-vercel-id": "iad1::iad1::cznpt-1788396828104-fe62f2166fbb" } } }

Happy to share more job ids or re-run a controlled case if it helps.

Thanks, Tadeu Tupinambá