Back to rankings

Gemini API: Gemini Omni 1.2 Flash Reaches General Availability

Gemini CLIView original changelog

Google released Gemini Omni 1.2 Flash (gemini-omni-1.2-flash) to general availability in the Gemini API. The model generates video up to 15 seconds, supports conversational editing with first-and-last-frame interpolation and per-turn video extension, and keeps characters and subjects more consistent across up to 6 reference images. Video output pricing drops 30% to $12.00 per 1M tokens, and paid-tier rate limits double.

Key Takeaways

  • Gemini Omni 1.2 Flash is now generally available, so developers can use gemini-omni-1.2-flash in production without preview labels.
  • Clips can reach 15 seconds, a longer single generation than earlier short-clip workflows.
  • First-and-last-frame interpolation generates 3 to 10 seconds of motion between two supplied frames.
  • Per-turn video extension adds 3 to 10 seconds each turn, up to 30 seconds total.
  • Up to 6 reference images improve character and subject consistency across generations.
  • Video output is 30% cheaper at $12.00 per 1M tokens, paid-tier rate limits doubled, and gemini-omni-1.1-flash auto-routes to 1.2 on December 3, 2026.

What Is New

Google made Gemini Omni 1.2 Flash generally available in the Gemini API under the model ID gemini-omni-1.2-flash. It is aimed at developers who generate and edit short videos programmatically.

Longer Clips and Richer Editing

The model can generate video clips of up to 15 seconds. Conversational editing now covers two workflows that matter for production use. First-and-last-frame interpolation lets developers supply a starting frame and an ending frame and have the model fill in 3 to 10 seconds of motion between them. Per-turn video extension adds 3 to 10 seconds to an existing clip on each turn, up to a total of 30 seconds.

Better Consistency

Google improved character and subject consistency across generations. The model now keeps a character or subject stable using up to 6 reference images, which helps when a project needs the same person or product to appear across several clips.

Lower Cost and Higher Limits

Video output pricing is 30% lower, at $12.00 per 1M tokens. Rate limits on paid tiers have doubled, so teams running batches of generations can send more requests before hitting a ceiling.

Migration From Omni 1.1 Flash

Google also announced a routing change. Starting December 3, 2026, requests to gemini-omni-1.1-flash will be routed automatically to gemini-omni-1.2-flash and billed at 1.2 Flash rates. Developers who pin the older model string do not need to change code to keep working, but should review output differences and cost before that date.