All model profiles

Model grammar profile · 3.1

Google Veo 3.1
Cinematography + subject + action + context + ambience

Apply Google’s five-part formula: cinematography, subject, action, context, then style and ambiance. Append dialogue and audio only when the selected version supports them.

Compile for Veo Official docs

Recommended sentence order

01cinematography→
02subject→
03action→
04context→
05style→
06dialogue→
07audio

Negative prompt policy

Use exclusions the way this surface expects.

When the target surface exposes a negative field, normalize exclusions to concise concepts rather than ‘no’ or ‘don’t’ instructions.

Typed parameters

Duration4 · 6 · 8
Aspect ratio16:9 · 9:16
Resolution720p · 1080p
Generate audioOn / off

Reliability constraints

  • API capabilities vary by Veo version, region, and Vertex/Gemini surface.
  • Timestamp blocks are guidance rather than a guaranteed parser contract.
  • Responsible AI filtering can block prompts or source images.

Source material

Google DeepMind Veo prompt guide Google Cloud Veo guide

Next profile

OpenAI Sora 2