Google puts Genie 4 on the Gemini API, at about ten cents a second on the pricing page
Thursday’s developer note says Genie 4 is available through the Interactions API and in Google AI Studio. It describes 720p video at 24 frames a second with native audio, and it does not print a rate. The pricing page does: $1.50 per million input tokens, $17.50 per million video output tokens, which Google converts to about $0.10 a second. The release notes call the model generally available. There is no free tier.
Short2 minIn-depth
James Okonkwo2 min read
- Genie 4
- Gemini API
- Pricing

Photo: GoogleSource page
Google’s developer note on Thursday is the distribution step Wednesday’s subscriber post did not take. The note, dated 19 November, says Genie 4 is now available through the Gemini API, via the Interactions API, and in Google AI Studio. A text prompt, a reference image, or both can start a world. The note describes continuous 720p video at 24 frames a second, with synchronized native audio, from ambient sound to character dialogue. Keyboard controls move the view. A text prompt typed during the session can change the weather, the lighting, the objects or the events. Gemini, the note says, keeps the world consistent. Every output carries a SynthID watermark. The note does not print a price. It points readers at a pricing page.
That page lists the model as models/genie-4, on the paid tier only. Input is $1.50 per million tokens, for text or a reference image. Text output is $9 per million tokens. Video output is $17.50 per million tokens, and the page says 48 kHz stereo audio is included in that video rate. The meter is 240 video tokens a frame, or 5,760 tokens a second at 24 frames a second. At $17.50 per million video tokens, the page converts that to about $0.10 a second, and about $6.05 a minute. It tells the caller to use complete() to end a session and stop the bill. Genie 4 is not on the free tier. It is not offered through the Batch API, or through Flex or Priority inference. Paid-tier prompts and outputs, the page says, are not used to improve Google’s products. The 19 November release notes call this general availability, and they add two details the blog post leaves out: the stream runs over WebRTC and WebSockets, and the live controls include WASD or the arrow keys as well as mid-session text. A maximum session length is in neither the blog post nor the pricing table.
The same pricing page lists a different video model with a different meter. Gemini Omni 1.2 Flash is about $0.07 a second of 720p, from $12 per million video tokens and 5,792 tokens a second, for generation and editing the page describes as up to 15 seconds. The Genie 4 announcement does not cite that line. The two rows are on one table. They are not the same product.


