r2v15sVirtual try-on, 15 seconds
One model photo plus front-and-back shots of three outfits become a single 15-second try-on sequence — 360° turns and hard match cuts, with identity, cut and drape held throughout.
Google Gemini Omni Flash1.0 → 1.1 → 1.2
One video family.
Three generations, one endpoint.
From the 1.0 original to 1.2's 15-second, multi-turn-editable clips — every Omni Flash generation ships on APIPod behind the same unified async video API. Pick a version by the job; switch by changing one model ID.
One shot through the PantheonReference-to-video · 15s · generated by Gemini Omni Flash 1.2
01 / 05 — The ladder
Pick the rung that fits the job.
Each generation widened what the fast tier can do. The newest isn't always the right tool — 1.0 still owns flexible 1–30s durations at a flat price.
I/O · May 2026
Gemini Omni
The original multimodal model: 1–30 second clips, up to 5 references, flat per-video pricing — still the cheapest way to make straightforward shots.
Open seriesGA · Aug 2026
Omni Flash 1.1
The fast tier grew 4K output, 7 references and 10-second-context scene extension; per-video pricing by resolution made 360p drafts nearly free.
Open seriesNov 2026
Omni Flash 1.2
Native 3–15 second clips, ten references, conversational editing and extension — per-second billing at 30% lower list pricing. The current default.
Open series02 / 05 — The matrix
Same bloodline, different limits.
The facts that decide a migration: durations, resolutions, reference counts, editing surface and how each one bills.
| Capability | Omni 1.0 | Flash 1.1 | Flash 1.2 |
|---|---|---|---|
| Model IDs | gemini-omni-{t2v,i2v,r2v,extend} | gemini-omni-flash-1.1-{t2v,i2v,r2v,extend} | gemini-omni-flash-1.2-{t2v,i2v,r2v,edit,extend} |
| Clip duration | 1–30s | 4–10s | 3–15s |
| Resolutions | 720p · 1080p | 360p · 720p · 1080p · 4K | 360p · 720p · 1080p · 4K |
| Reference images | ≤ 5 | ≤ 7 | ≤ 10 |
| Editing surface | Extend | Extend | Multi-turn edit + extend |
| Audio | Native, synced | Native, synced | Native, clearer lip-sync |
| Billing model | Flat per video | Per video × resolution | Per second × resolution |
On APIPod every generation returns the same task_id + polling contract. Google auto-routes 1.1 Flash traffic to 1.2 from December 3, 2026.
03 / 05 — 02 — The reel
Eight more clips, one model.
What the newest generation looks like in practice — every clip below is a single Omni Flash 1.2 generation (or one multi-turn chain), straight from Google's launch material.
r2v15s
i2v10sFirst-to-last frame, no drift
An exterior start frame and an interior end frame join in one continuous camera push — roof cats, ladle and signage all survive the trip.
edit15sMulti-turn background swaps
Turn-by-turn prompts swap the backdrop — striped tent, oak forest, stable, market, castle — while both performances, armor reflections and framing stay put.
edit15sMotion transfer across turns
Turn 1 generates a rippling octopus; turn 2 replaces it with a dancer whose gown mirrors the exact same fluid motion.
r2v15sA 15-second science explainer
One style frame becomes a 4-shot macro explainer with accurately spelled, motion-tracked callouts — egg, larva, queen — inside the hive.
r2v15sTypographic motion poster
A reference poster drives a vertical animated build — HALO, GLOW, LUMEN — locking materials, shadows and letterforms to the final frame.
r2v15sFour voices, the right faces
A conservatory scene image plus character, mouth and one-second voice references: each sampled voice lands on the right character with expressive lip-sync and comedic timing.
t2v14sCrash zoom out
Text only: an intimate close-up on a castaway pulls back in one continuous move to reveal the whole island.
Clips from Google's Gemini Omni 1.2 Flash announcement, compressed and re-hosted on APIPod's CDN. Originals on blog.google.
04 / 05 — Moving to 1.2
Three steps, one field.
Google routes 1.1 Flash requests to 1.2 from December 3, 2026 — and because the request shape never changed, migrating is mechanical.
Swap the model ID
- "model": "gemini-omni-flash-1.1-t2v"+ "model": "gemini-omni-flash-1.2-t2v"
Change gemini-omni-flash-1.1-t2v to gemini-omni-flash-1.2-t2v (likewise -i2v / -r2v). Payload fields you already send — prompt, image_urls, duration, resolution, aspect_ratio — carry over unchanged.
Relax the duration
- "duration": 10 // 4–10+ "duration": 15 // 3–15
1.2 accepts any whole second from 3 to 15 on t2v, first-frame and reference jobs. Keyframe (first+last) clips still cap at 10s; extend adds 3–10s; edit ignores duration and follows its source.
Expect per-second bills
billed = seconds × resolution (edit: source s · extend: source + added s)
Billing moves from per-video to per-second × resolution, so short clips get cheaper and long ones cost what they should. Editing bills the source video's seconds and extension bills source plus added seconds — budget accordingly.
05 / 05 — FAQ
Version questions
Which Omni Flash version should I use?
Default to 1.2: longest clips, most references, and editing plus extension. Use 1.1 if you depend on per-video pricing or its exact request routing. Use 1.0 when you need flexible 1–30 second durations at a flat per-video price.
Gemini Omni Flash 1.2 vs 1.1 — what actually changed?
Native clips grew from 10 to 15 seconds with per-second duration control; references doubled from 7 to 10 images; conversational multi-turn editing and extension arrived; speech sounds more natural with clearer lip-sync; and list pricing dropped about 30%. In Google's side-by-side studies 1.2 won every reported metric.
Is 1.1 going away?
Google auto-routes 1.1 Flash requests to 1.2 starting December 3, 2026. On APIPod the 1.1 model IDs keep answering (with the same request shape), so nothing breaks — but new features and price improvements land on 1.2.
What is Gemini Omni Flash 1.2 Lite?
A lower-priced tier of the same Omni Flash 1.2 model, offered as gemini-omni-flash-1.2-lite-{t2v,i2v,r2v,edit,extend}. The request shape is identical to 1.2, so switching between them is a model-ID edit.
Do all versions generate audio?
Yes — synchronized native audio (dialogue, ambience, score) is a family-wide capability. 1.2 adds more natural speech, clearer lip-sync and distinct multi-speaker conditioning.
How is each version priced on APIPod?
Omni 1.0 bills a flat rate per video. Flash 1.1 bills per video by resolution. Flash 1.2 bills per second by resolution. On 1.2, editing bills the source video's seconds and extension bills source plus added seconds. See each series page for live rates.
Do I need different code for each version?
No. All three generations POST to /v1/videos/generations and return a task_id you poll at GET /v1/videos/status/:task_id. Switching generations or tasks is a one-field model-ID change.
One key, every generation.
Start on 1.2 today; 1.1 and 1.0 stay available on the same key for whatever your pipeline demands.