Gemini Omni Flash Video: Token Pricing Is Not a Flat Per-Second Fee

Two input/output modalities in one table

Google lists the developer identifier gemini-omni-1.1-flash with Standard API input pricing of $1.50 per million tokens across listed input modalities. Output text is $9 per million tokens and output video is $17.50 per million output tokens. Those are not the same unit as a Veo per-second list rate. Model the media that actually enters and leaves the service, including source frames, supplied audio and generated tokens, rather than assigning all data to one blended input/output pool.

Where the roughly ten-cents-a-second estimate comes from

The vendor states that 720p output consumes 5,792 video tokens per second. Multiplying 5,792 / 1,000,000 × $17.50 gives approximately $0.10136 per second of generated video, often rounded to about $0.10. That is a video-output-only mathematical conversion. Input and other output types remain additional. Nothing in this ratio guarantees a constant token-per-second rate at other resolutions or future model versions.

Why this differs from Veo credits

A token-output service, a per-second Veo API rate and a consumer subscription measured in credits are three different billing systems. A video editor may see 720p clips in all three but still face different queue terms, prompt limits, effects, editing features and ownership conditions. Do not convert consumer plan credits into API cash without an explicit provider rate and eligibility rule. Keep model_version and unit fields together in price_snapshots.csv to avoid a misleading comparison table.

Plan for edits and multi-clip outputs

For a purely hypothetical 60 seconds of billed 720p Omni video output at the documented density, 347,520 video tokens × $17.50 per million is $6.0816 before input and text output. If an editing task produces two revisions that are both billable, the usage may increase independently of the final accepted duration. Store total observed generated-video tokens from the API response rather than reverse-calculating everything from completed timeline seconds.

Before choosing a paid tier

Obtain the billing quote for the exact Google product, account geography, tier, model or subscription, currency and commitment length. If the provider offers several kinds of meters, forecast each independently rather than summing quantities in incompatible units. The relevant captured meters here include input_multimodal_tokens_per_1m, output_text_tokens_per_1m, output_video_tokens_per_1m. Build a low, expected and high workload from your own request logs or editorial schedule; a pricing page is not a credible estimate of how many retries or approvals your team will need.

For governance, add an approval threshold for paid overages, reconcile invoices with metered activity, and schedule a quarterly or release-triggered source audit. The absence of an identified enterprise rate, promotional expiry or regional tax line is not evidence that the amount is zero. This guide does not represent a product trial, capability benchmark, procurement offer or warranty of uninterrupted access.

Historical availability and boundaries

Preserve the precise source version, checked date and any stated effective date before reusing these figures. A model launch announcement may confirm an introductory token rate without proving which private offers remain available; a mutable plan page can describe a later product state. If a source changes after this cutoff, publish a new snapshot instead of silently editing the historical record. Specialised product evaluations belong to the relevant AI 02–09 editorial teams, while this article documents the billing model and its limitations.