What changed in 1.2 Flash

Google made Gemini Omni 1.2 Flash generally available on November 5, 2026. This is a production update to an existing video model, not a new foundation model — the pitch is volume work (product clips, virtual try-ons, ad variants) rather than one cinematic hero shot. Four changes matter:

  • Longer clips. Native generation and editing now reach 15 seconds, up from 10 seconds in Omni 1.1. Duration is selectable in one-second steps from 3 to 15 seconds, so a full social spot fits in one generation.
  • 30% cheaper at 720p. The published API price is $0.07 per second of 720p video, down from Omni 1.1 Flash; Google says other resolutions see similar savings.
  • Multi-turn conversational editing. Refine the same clip — swap a product, change lighting or weather, switch aspect ratio — instead of regenerating from scratch. It works across the full 15-second clip, which is the real cost lever for ad variants.
  • Stronger reference adherence. Combine character, product or style images, motion clips up to three seconds, and audio cues; faces, garments, packaging typography and brand colors hold across camera angles. Audio references are currently allowlist-gated.

Also in the release: a draft-then-upscale workflow (draft at 360p or 720p, fix details with follow-up edits, then upscale only the keeper to 1080p or 4K), more natural dialogue with better lip-sync and multi-speaker conditioning, and an imperceptible SynthID watermark on every output. Google Cloud backs it with Provisioned Throughput for enterprise workloads.

How it compares: length, price, control

ToolNative clipPublished 720p priceEditing modelAccess
Gemini Omni 1.2 Flash3–15 s, 1 s steps$0.07/s, -30% vs Omni 1.1 (Google official)Multi-turn conversational edit on the same clipGemini API, AI Studio, Cloud Console, Flow, Google Vids, Gemini app (AI Ultra), Google Ads
Kling 4.0 / 3.0~10 s per generation, extendableKling's own API: 1.0 credit/s at 720p, no official USD/s rate. Runway Dev lists Kling 3.0 Pro at $0.12/s silent / $0.17/s with audioReference-to-video, video editing and extension; revisions mostly mean re-generatingKling app + Kling API; also via Runway Dev
Runway Gen-4.5~10 s~$0.12/s (12 credits/s at $0.01/credit, third-party compilation)Timeline-based editor, professional motion controlsRunway app + Dev API
Sora (OpenAI)~10–20 sSubscription-led; a historical Sora 2 API rate of $0.10/s at 720p is quoted by third parties, and API availability has changedRe-prompt / storyboard, no native multi-turn clip editingChatGPT app, API where available

Read the price column carefully — the units do not match. Google bills dollars per second of output. Kling's own API bills credits per second, and when a video is supplied as input, the billable total duration can be computed as output x2 or input + output. Runway Dev bills credits at $0.01 each, with audio on/off shifting the rate by roughly 30–40%. Recompute against your own duration, resolution and audio settings on each vendor's pricing page before budgeting.

Two more limits on this table: no independent quality leaderboard covers Omni 1.2 Flash yet (it shipped days ago), so consistency and reference-adherence claims are Google's own plus launch-partner quotes from Artlist, FLORA, invideo, Krea and Picsart. Third-party comparison sites also disagree with each other and occasionally with official docs — one claims 60–90 second Omni clips, which contradicts Google's 15-second cap. Treat the table as a starting point, not a quote.

Where to use it: API and product entry points

  • Developers: generally available on the Gemini API and Google Cloud; use Google AI Studio to prototype and read the docs and pricing, and the Cloud Console for enterprise setup.
  • Teams without code: Google Vids (15-second generation inside Workspace; rollout can take 1–3 days depending on the release track), Flow for directing scenes and multi-turn edits, Gemini for business, and Google Ads for scaled ad variants.
  • Consumers: rolling out in the Gemini app to Google AI Ultra subscribers.

A cheap first test of the workflow that justifies the price cut:

  • Generate a short draft at 360p or 720p in AI Studio.
  • Send a follow-up instruction in the same session — change the background, lighting or aspect ratio — and check that the rest of the clip stays intact.
  • Upscale only the take you keep to 1080p or 4K. Draft cheap, edit instead of regenerating, render high-res once.

Caveats and sources

  • Google Cloud blog, "Gemini Omni 1.2 Flash: Fast, cost-effective, and reliable video generation and editing now generally available" — GA, $0.07/s at 720p, 15 s, -30%, SynthID, Provisioned Throughput.
  • Google blog, "Gemini Omni 1.2 Flash: Fast and cost-effective with longer clips" — 3–15 s in one-second steps, multi-turn editing, reference adherence, entry points.
  • Kling official API pricing (klingai.com/dev/pricing) — credit-per-second rates at 720p; Runway Dev, "Kling 3.0 API: access, pricing and code" — $0.12–0.17/s, prices stated current as of November 6, 2026.
  • Third-party pricing compilations for Runway Gen-4.5 and Sora 2 (apiframe.ai, tech-insider.org) — not independently verified against vendor pricing pages at publish time.

← Back to all guides