Skip to content

Google Veo 3.1

Overview

Veo 3.1 is Google’s newer Vertex AI video line. Google’s official model page and prompt guide frame it as a more controllable, more production-oriented successor with better prompt adherence, improved audiovisual quality, and support for workflows such as first-and-last-frame conditioning and reference-image guidance. It is the right model family to document when teams care about cinematic direction and creative control rather than only simple prompt-to-video conversion.

API base URL

https://us-central1-aiplatform.googleapis.com/v1

Example payload

{
  "instances": [
    {
      "prompt": "Slow cinematic drone shot over a futuristic skyline",
      "duration": "8s",
      "aspectRatio": "16:9",
      "fps": 24,
      "resolution": "1080p",
      "sampleCount": 1
    }
  ],
  "parameters": {
    "aspectRatio": "16:9",
    "durationSeconds": "8s",
    "resolution": "1080p",
    "safetySetting": "block_only_high",
    "personGeneration": "allow_all",
    "includeRaiReason": true
  }
}

Our Current Pricing

Veo billing is implemented in the video workflow rather than being read from the provider response. The code computes cost from duration, model line, resolution tier, quality multiplier, and requested video count. For Veo 3.1, the workflow maps veo-3.1-generate to 0.40 USD/sec by default or 0.60 USD/sec at 4K, and veo-3.1-fast to 0.15 USD/sec by default or 0.35 USD/sec at 4K.

Quality multipliers are:

draft    -> 0.5
standard -> 1.0
high     -> 1.5
ultra    -> 2.0
unknown  -> 1.0

Final formula:

total_cost_usd =
  duration_seconds *
  base_cost_per_second(model, resolution) *
  quality_multiplier(quality) *
  video_count

Final rounding is to 6 decimal places.

The model record itself also contains a simpler per-second price map with an audio modifier, but MODEL_COSTS_OVERVIEW.md is explicit that the effective workflow formula is the one above. For billing, prefer the workflow logic because that is what the application actually executes.

The best calculation strategy is request-driven. Store model, duration, resolution, quality, and sampleCount or requested count before submission, then compute cost deterministically once the request is accepted. If the provider later returns richer usage or audio-specific metering, you can compare it for reconciliation, but it is not the current source of truth.

Sources