Field Guide
Gemini Omni Prompt Guide
A practical field guide. What works, what breaks, what to avoid. Based on DeepMind's official prompt guide + community testing from PixVerse, Atlas Cloud, Chrome Unboxed, and Medium's Gemini Omni Prompt Playbook.
What changed in the last 30 days
A dated timeline of every material change since launch. We update this whenever something shifts.
- 2026-05-19 — Gemini Omni Flash launches at Google I/O 2026. DeepMind publishes the model card 🟢. Rolls out same day to AI Plus / Pro / Ultra subscribers globally via the Gemini app + Google Flow, and free in YouTube Shorts / YouTube Create App (18+). API for developers and enterprise promised "in coming weeks." (Source: blog.google 🟢.)
- 2026-05-19 — Google restructures AI subscription pricing: Plus $7.99, Pro $19.99, new Ultra tier $99.99 (creator), existing Ultra reduced $250 → $200. (Source: Google One subscriptions blog 🟢.)
- 2026-05-19~20 — Avatars feature begins rolling out to paid subscribers. 18+, geo-blocked in EEA / Switzerland / UK. (Source: Google Gemini Avatar Help 🟢.)
- 2026-05-20 — Community reports widespread false-positive "policy violation" rejections on harmless prompts. Google VP Josh Woodward publicly acknowledges and says Google is investigating. Status as of 2026-05-29: fix not announced. See "the two bugs" section below.
- 2026-05-28 — Woodward announces fixes to a different bug — the compute/quota one. Failed requests no longer count against quota; AI Ultra Omni allowance doubled; Gemini 3.1 Flash-Lite prompts now free; Pro per-prompt quota capped. (Sources: 9to5Google 🟡, The Hans India 🟡.)
Source tier: 🟢 official Google/DeepMind · 🟡 reputable hands-on / news · 🔴 unverified.
Who can use Gemini Omni Flash (and what it costs)
Access and pricing checked against the official sources below. The launch timeline, historical tests, and other sections retain their original dates.
Google Flow access and credits
Google Flow provides 50 daily credits with or without a subscription. Free video generation may be unavailable during peak hours (roughly 2–5 PM UTC). Subscribers receive additional monthly Flow credits:
| Plan | Additional Flow credits / month |
|---|---|
| Google AI Plus | 200 |
| Google AI Pro | 1,000 |
| Google AI Ultra $100 | 10,000 |
| Google AI Ultra $200 | 25,000 |
Source: Google Flow credit help 🟢. These are Flow credit allowances, separate from Gemini app usage limits and Gemini API billing.
Omni Flash credit cost in Google Flow
| New video length | 720p | 360p |
|---|---|---|
| 4 seconds | 7 | 4 |
| 6 seconds | 10 | 5 |
| 8 seconds | 12 | 6 |
| 10 seconds | 15 | 7 |
Values are Flow/AI credits per generation, not per request; a request can produce multiple videos. Editing an uploaded or generated video costs 40 credits at any length or resolution. Source: Google's per-generation credit table 🟢. Costs can change; check Flow's prompt Settings before generating.
Historical observations (May 2026)
These launch-period observations are retained for context and do not represent current pricing:
- Our own testing (Cited, 2026-05-28): a fresh 10-second 16:9 generation cost 30 credits. A single conversational edit on the same clip cost 40 credits — i.e., editing is 33% more expensive than re-generating from scratch. 🟡
- WaveSpeed's launch coverage reports a 4-second fresh clip at 15 credits plus 40 for an edit — different generation cost, same edit cost. Could be a shorter-clip discount, or pricing has shifted since. 🟡
Cost takeaway: a new clip at the listed resolutions costs fewer credits than a 40-credit edit. Choose whether to regenerate or edit based on what you need to preserve in the existing clip.
Where to use it
- Google Flow — video generation and conversational editing; credit costs are listed above.
- Gemini API —
gemini-omni-1.1-flashis generally available on the paid tier; unavailable on the free tier. Standard API pricing is $1.50 per million input tokens, $9.00 per million text output tokens, and $17.50 per million video output tokens. At 720p, 5,792 output tokens per second makes video output approximately $0.10/second; input and any text output are billed separately. Flow subscription credits do not describe these API charges. - Gemini app and YouTube launch access is recorded in the May timeline above; their current availability and limits were not rechecked in this section.
The base formula
Most working Gemini Omni prompts follow this structure:
[Subject] + [Action] + [Setting] + [Camera] + [Lighting] + [Style] Example: "A red panda chef tossing pizza dough, in a cozy mountain kitchen, low-angle close-up, warm tungsten light, Pixar-style 3D animation".
This is the same DNA as Sora / Veo prompts, but Omni adds two unique elements that most other models don't have:
- The "
@username" character summon syntax (for Avatar feature) - The conversational editing chain ("now change X, keep Y identical")
The opening-line trick
Lock these three things in the very first sentence:
Create a [duration]-second [aspect-ratio] [genre] video in one continuous shot. Why: Omni interprets "one continuous shot" as "no cuts" and respects the time / aspect ratio you specify upfront. Specifying these inline beats putting them in metadata.
Camera vocabulary Omni understands
From DeepMind's prompt guide, these terms are explicitly parsed:
Camera motion verbs
- Push: push in / punch in / dolly zoom
- Pull: pull-back / pull-up / pull-down
- Reveal: ascend revealing / pull-back and rotate
- Orbit: orbit around / sweep around / circle
- Pan: pan left / pan right / vertical pan
- Static: locked off / fixed / oner / continuous shot
Style references
- natural smartphone zoom
- film camera (warm, slight grain)
- webcam style (compressed, slightly soft)
- handheld (micro-jitter, organic)
Duration sweet spots
| Goal | Best duration | Aspect ratio |
|---|---|---|
| Mood / cinematic | 8-10s | 16:9 or 2.39:1 |
| Product hero | 6-8s | 1:1 or 16:9 |
| Reels / TikTok / Shorts | 5-7s | 9:16 |
| Slow-motion impact | 5-6s | 1:1 or 16:9 |
| Timelapse | 10s (max) | 16:9 |
| Avatar talking head | 10s (max) | 16:9 or 9:16 |
Gemini Omni Flash hard limit: 10 seconds per clip. Source: TechCrunch launch coverage.
The "Keep X identical" lock
When using conversational editing (Omni's flagship feature), every follow-up turn should explicitly list what to preserve. Pattern:
[Change instruction]. Keep [X, Y, Z] exactly the same. Without this lock, Omni may re-style the entire scene when you ask it to change one element — losing the consistency that's the whole point of conversational editing. (Documented by Atlas Cloud's hands-on testing.)
The trigger pattern (for VFX)
One of Omni's strongest patterns — used in Google's own viral demos (mirror-arm-transformation, bubble sculpture, origami ships):
[Base scene]. When [specific trigger action], [specific transformation]. Keep [list] identical. Example: "A woman reaches toward a mirror. When her fingertips touch the glass, make the mirror ripple like liquid and her arm turn to reflective mirror material. Keep the parlor and lighting identical."
"It's blocking my prompt" — the two bugs you need to keep straight
Since launch, the phrase "Gemini Omni is blocking my prompts" has been used to describe two completely different bugs. Mixing them up is the most common source of confusion in the community right now.
Bug A — The false-positive policy block (still unresolved as of 2026-05-29)
Omni rejects clearly harmless prompts (Cat walks between rabbits,
Change the background to night) with a generic "I can't generate that video" / "policy violation" message —
claiming a rule violation that doesn't exist. Widely reported from 2026-05-19 onward. Google VP Josh Woodward
acknowledged the issue publicly and said Google was investigating and collecting affected examples.
As of today, no fix has been announced; users are still posting fresh examples on the
Google AI Developers Forum 🟡.
For the full breakdown of what's a real policy vs. this bug, see our
deep dive on Bug A.
Bug B — The quota counting bug (fixed 2026-05-28)
Unrelated to the policy block — this was a billing bug where failed generations (including ones killed by Bug A) were still being deducted from your quota. On 2026-05-28, Woodward announced:
- Failed requests no longer count against quota.
- AI Ultra Omni allowance doubled.
- Gemini 3.1 Flash-Lite prompts now completely free.
- Pro per-prompt quota capped to prevent runaway consumption.
Source: 9to5Google 🟡.
How to tell which one hit you
- You got a "policy violation" / "I can't generate that video" message on an obviously innocent prompt → Bug A. Don't rewrite the prompt 10 times — it's the bug. Try again later or rephrase slightly.
- Your quota dropped despite a failed / refused generation → Bug B. That should be fixed now; if you still see it, contact Google support with the request ID.
Known failure modes (be honest about these)
DeepMind's own model card explicitly lists these as "Known Limitations" 🟢 — so when you hit them, it's not a prompt-writing problem, it's a model boundary.
Text rendering
Any onscreen text — labels, signage, captions, brand logos — degrades. Avoid mentioning text overlays in your prompt. Verified by PixVerse hands-on.
Hand articulation
Hands holding objects, sign language, typing — fine articulation drifts. Frame to hide hands when possible, or accept some imperfection.
Multi-shot character consistency
Per Atlas Cloud's multi-turn review:
Omni scores 3/5 on character consistency across 4+ shots. Use @character_name
with a reference image for best results, and accept drift past shot 4.
Complex motion
Per digit.in's test: complex actions (dancing, gymnastics, instrument playing) show AI artifacts more than static shots. Simple actions (walking, standing, talking) work best.
Word count over 50
Per Seaart's analysis: prompts longer than ~50 words dilute focus and reduce output quality. Be specific but concise.
What NOT to do
- Don't write Veo-style adjective stacks — DeepMind explicitly says Omni "doesn't need to be as prescriptive as Veo". Natural language beats template formulas.
- Don't change multiple variables per turn — split into separate conversational turns.
- Don't reference copyrighted IPs by name — "Studio Ghibli style" is risky; "hand-painted watercolor animation" is safer.
- Don't use Veo / Sora-specific syntax — Omni parses some of it but optimizes for its own conversational style.
- Don't request hardware brand names ("shot on DJI Mavic Pro") — Omni doesn't parse camera brands; use motion verbs instead.
Avatar feature hard rules
- Age requirement: 18+
- Geography: NOT available in EEA / Switzerland / UK at launch
- Language: English only at launch (May 2026)
- Watermark: every video carries Google SynthID, non-optional, embedded in pixels
- Reference recording: clear eyes / nose / mouth, no sunglasses / masks / hats covering face, no other faces in background
Source: Google Gemini Avatar help page.
Sources
- Official: DeepMind Gemini Omni Prompt Guide
- Official launch: blog.google announcement
- Avatar specifics: Google Gemini Avatar Help
- Hands-on: Atlas Cloud Features Overview
- Hands-on: PixVerse Model Review
- Avatar hands-on: Chrome Unboxed Avatar Test
- Playbook: Medium — Gemini Omni Prompt Playbook
- Limitations review: digit.in hands-on
- Model card (official): DeepMind — Gemini Omni Flash
- Pricing restructure: Google One AI subscriptions blog
- Quota fix announcement: 9to5Google — new usage limits
- Active bug thread: Google AI Developers Forum #147152
Last updated 2026-05-29. We re-verify every claim and update this page whenever Google ships a change. Spot something wrong or outdated? Tell us — we read every email.