LumeAPILast verified: July 25, 2026
Choose Veo 3.1 Fast or Quality when the ad needs a fixed eight-second, high-fidelity audiovisual production candidate with Google-documented text/image video generation and native audio. Choose Grok Imagine 1.5 when the application already has an approved source image and needs an economical 6-to-30-second image-to-video test lane. Do not treat Grok as a like-for-like Veo text-to-video substitute: xAI's current native model page lists Grok Imagine Video 1.5 as image-to-video, while the LumeAPI catalog's route notes should be verified in staging for the exact request shape you plan to send.
This article answers one decision: which route should a social-ad workflow test when it must turn a source asset into a reviewed short video? The direct answer is conditional because input mode, duration, and required quality change the economics more than the model label does. No paid ad campaign or billable head-to-head generation was run for this page.
Direct routing answer
| Ad workflow condition | Start with | Why | Boundary |
|---|---|---|---|
| Premium eight-second audiovisual concept, text/image direction, and a selected brand final | Veo 3.1 Fast for an iteration lane; Quality/standard for selected finals | Google documents 8-second outputs, native audio, and text/image guidance for Veo 3.1 | Final rate depends on tier, resolution, audio, and input configuration |
| Existing approved product/character image needs an affordable motion test | Grok Imagine 1.5 | LumeAPI lists $0.03/s, 6–30s, audio included, and image_urls support where enabled | Native xAI docs list image-to-video, so require the source image and verify the gateway field contract |
| No source image; the job is text-to-video | Veo first | Google's documented Veo contract includes text input | Do not infer Grok 1.5 native text-to-video availability from another Grok route or a gateway label |
| High-resolution 1080p/4K brand final | Veo standard after a lower-cost candidate passes | LumeAPI lists explicit Veo resolution/audio tiers | Never compare a Grok base 480p card rate to a 4K Veo final |
Input contract and current LumeAPI USD rates
| Route | Exact LumeAPI ID | Deployable catalog signal | Native provider context | LumeAPI price |
|---|---|---|---|---|
| Veo 3.1 Lite | veo3.1-lite | Fixed 8s, T2V only, no image URLs | Google documents a Lite variant; verify exact LumeAPI parameter support | from $0.03/s |
| Veo 3.1 Fast | veo3.1-fast | Fixed 8s, T2V/I2V, 720p/1080p/4K tiers | Google documents Veo Fast for speed/business use and native audio | from $0.05/s |
| Veo 3.1 Quality | veo3.1-quality | Fixed 8s, higher-fidelity tier, T2V/I2V | Google documents 8-second high-fidelity Veo with audio | from $0.20/s |
| Veo 3.1 standard | google/veo-3.1 | 1080p/4K and audio price bands | Google documents text/image inputs and native audio; feature availability depends on mode | $0.20–$0.60/s |
| Grok Imagine 1.5 | grok-video | 480p/720p, 6–30s, audio included, image_urls where enabled | xAI documents grok-imagine-video-1.5 as image-to-video, $0.08/s at 480p plus image input billing | from $0.03/s |
All values are USD and were checked July 25, 2026. The LumeAPI rate is the actionable card price for a LumeAPI request. xAI's native $0.08/s 480p figure is provider context, not a promise that every gateway request is feature- or configuration-equivalent; xAI also bills image input separately. Google documents Veo's native audio and fixed eight-second output, but a mode's final bill depends on the chosen resolution and audio state.
Same-job cost: an eight-second vertical ad
| Configuration | LumeAPI rate | Eight-second generation cost | Eligible job |
|---|---|---|---|
| Grok Imagine 1.5 base | $0.0300/s | $0.2400 | Source-image motion test; verify actual eligible resolution and input contract |
| Veo 3.1 Lite base | $0.0300/s | $0.2400 | Text-only concept where image input is unnecessary |
| Veo 3.1 Fast base | $0.0500/s | $0.4000 | Faster Veo iteration candidate |
| Veo 3.1 Quality | $0.2000/s | $1.6000 | Selected premium eight-second candidate |
| Veo 3.1 standard 1080p with audio | $0.4000/s | $3.2000 | High-resolution audiovisual final |
The $0.24 Grok and Lite cards are not interchangeable. Grok's native 1.5 documentation says image-to-video, while Lite is explicitly text-to-video only in the LumeAPI catalog. In a real ad workflow, input availability is a release condition, not a minor parameter.
For an approval-cost illustration, if a $0.24 Grok candidate is approved 45% of the time, its generation cost per approved clip is $0.533. If a $1.60 Veo Quality candidate is approved 85% of the time, its cost is $1.882. The lower-cost route remains cheaper in this example, but a creative team may still select Veo for a high-value launch when the manual-rework or brand-risk difference is large. These are transparent assumptions, not benchmark results.
Application workflow: approved still to reviewed vertical ad
- In a DAM, Shopify, or content-management system, identify 20 approved product/character stills. Store the source asset ID, allowed claims, target platform, aspect ratio, duration, and whether native audio is required.
- Route jobs with an approved source still to Grok Imagine 1.5 only after a staging request confirms the current LumeAPI
image_urlsbehavior. Route text-to-video briefs to Veo; do not quietly replace a missing source image with a different model. - Submit via
POST /v1/videos, persist a job record, and pollGET /v1/videos/{id}. The UI must handle queued, failed, expired, and cancelled jobs without duplicate submits. - In CapCut, Adobe Premiere, or an ad-review queue, score each completed clip on source-asset fidelity, motion coherence, audio usability, platform-safe crop, brand/legal copy overlay space, and whether a human can approve it without material rebuild.
- Log model, input mode, input-media count, duration, resolution, audio, job ID, billed rate, retry count, completion time, and
approved. Compare cost per approved clip by workflow, not only by model.
This method prevents a common false comparison: evaluating Veo text-to-video against Grok image-to-video with unrelated prompts, then declaring a model winner. The source asset and intended placement must be held constant for a useful decision.
When Veo 3.1 is the right first route
Choose Veo when your social-ad workflow needs a fixed 8-second production candidate with text or image guidance, native audio, and a documented premium ladder. Google explicitly identifies Veo 3.1 as an 8-second, high-fidelity video model and describes image guidance, native audio, portrait/landscape framing, extensions, and frame-specific generation in its native API. That makes it the safer first candidate for a text-to-video brief or selected brand final, subject to verifying the current LumeAPI route's supported fields.
Use Lite or Fast to validate a visual direction before paying Quality or standard rates. Do not send every concept directly to 4K/audio tiers.
When Grok Imagine 1.5 is the right first route
Choose Grok Imagine 1.5 when the ad begins from an approved image and the team needs many economical motion alternatives. The route's listed $0.03/s base rate, 6–30 second duration, and included-audio catalog note make it a strong test candidate. The required caveat is material: xAI currently documents this specific native model as image-to-video and lists separate source-image input cost. Build the application around a source-image workflow; do not promise text-to-video unless the exact LumeAPI route is verified in staging.
The creative gate remains strict: a low-cost video that alters a product, subject, or claim is not a cheaper ad. Keep human review for identity, likeness, moderation, and public-claim risk.
Final recommendation
For social video ads, choose Veo 3.1 when the job starts from text or needs a selected premium audiovisual eight-second final. Choose Grok Imagine 1.5 when you already have an approved source image and want low-cost 6–30-second motion exploration. Make source-input mode an explicit route requirement, store the exact tier/resolution/audio configuration, and select the production default from approved-ad cost and brand-blocker rate.
Related reading: AI video model comparison, multi-model API, and usage logs.
Sources and method
- Google Veo 3.1 API guide and Gemini API pricing - native Veo input, audio, duration, and price context.
- xAI Grok Imagine Video 1.5 and xAI pricing - current native image-to-video and price context.
- LumeAPI model catalog - exact gateway IDs and USD rate notes checked July 25, 2026.
No billable inference was run for this article. Its original contribution is the input-mode decision rule, same-job price matrix, and source-image-to-approved-ad workflow.