Google DeepMind's Gemini Omni 1.2 Flash is now available on Design Arena, the AI model evaluation and creation platform, bringing native video generation of up to 15 seconds along with multi-turn conversational editing and improved reference following.
Gemini Omni 1.2 Flash · Google DeepMind
Google DeepMind's Gemini Omni 1.2 Flash arrives on Design Arena with 15-second native video and conversational editing
The model is now accessible on the Design Arena platform, offering native video clips of up to 15 seconds, follow-up edits over successive turns, and improved reference following.
15 s
Maximum native video length per generated clip
Multi-turn
Refine a clip with follow-up instructions instead of re-prompting from scratch
Length ceiling, drawn to scale
Each block is one second. The full stack is the 15-second maximum.
0 s 15 s maximum
Takeaway: 15 seconds is a short-form ceiling, enough for a brief clip but not long-form video.
How conversational editing works
Prompt
→
Native clip up to 15 s
→
Follow-up edit
→
Refined clip
Edits repeat over successive turns, adjusting individual details rather than restarting the prompt.
What the update offers
Native video generation up to 15 seconds
Multi-turn refinement of generated clips
Improved adherence to supplied reference material
A way for creators to test it alongside other systems
What remains unestablished
Independent benchmark results
Detailed user evaluations
Performance versus competing video generators
Pricing, exact resolution, and availability limits
What to watch
Will the 15-second ceiling and editing features hold up under broader testing?
Broader testing results
Whether availability expands beyond the current platform
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article...