Google Rolls Out Three Separate AI Products: Gemini Omni Flash API, Nano Banana 2 Lite Image Model, and Gmail Live Voice Search Beta

Google Rolls Out Three Separate AI Products: Gemini Omni Flash API, Nano Banana 2 Lite Image Model, and Gmail Live Voice Search Beta

Google has added three more products to its commercial stack in quick succession.

Gemini Omni Flash Hits the API

Gemini Omni Flash debuted for consumers at Google I/O 2026. The model is now rolling out to developers and enterprise customers through an API, according to VentureBeat, which previously noted that without a programmatic interface, Omni was a consumer and prosumer tool, not a production one. This API rollout changes that.

The pitch is unification. Most enterprise teams building AI video today are stitching together multiple separate tools: an LLM for the script, a text-to-image model, an image-to-video model, a lip-sync tool, and a voice generator. Each carries its own vendor contract, billing structure, and data-handling policy. Omni Flash collapses that stack into one model that accepts text, images, and existing video clips and returns a finished clip with synced audio.

The more commercially interesting feature is conversational editing. Rather than regenerating a clip from scratch when a legal team wants one line changed, a user can issue a follow-up instruction that builds on the previous state of the clip. Relight a product shot, reframe it, change on-screen text without losing everything that already worked. VentureBeat describes it as "the difference between booking a reshoot and sending a note."

Reference-image support adds another layer. Feed the model a photograph of a specific product, logo, or location, and it carries the real object's coloring and rough shape into the generated scene rather than inventing a generic stand-in. VentureBeat notes the match is not pixel-perfect but close enough for recognition.

The fair counterargument from enterprise buyers is real: API access does not equal production readiness. Google's own model card is candid that holding consistency across edits and rendering accurate text remain open problems. Clips currently cap at 10 seconds. Google has not published complete per-minute video pricing, which makes total cost-of-ownership comparisons difficult. Every Omni clip carries Google's SynthID watermark, and Google is extending C2PA Content Credentials across its generative tools.

Nano Banana 2 Lite: 4-Second Images at $0.034 per Thousand

On the image side, Google also released Nano Banana 2 Lite, officially designated Gemini 3.1 Flash-Lite Image on the API, according to VentureBeat. It is available immediately through Google AI Studio, the Gemini API, and the Gemini Enterprise Agent Platform.

The core numbers: standard 1k resolution images in under four seconds, priced at $0.034 per 1,000 images. Google's internal benchmarks put its Text-to-Image arena Elo score at 1,251, which VentureBeat says surpasses the previous Nano Banana 1 score of 1,151 by a meaningful margin. The model also edges out the bulkier NB Pro, which sits at 1,245 in the same text-to-image track.

The limitations are spelled out in Google's own technical documentation. NB2 Lite is capped at 1k resolution. The broader NB2 and Nano Banana Pro lines support 1k, 2k, and 4k outputs. If your workflow needs anything above 1k, this model is not it.

Where the product is competitive is throughput. For automated asset generation at high volume — catalog images, social graphics, data visualization drafts — the sub-four-second generation and flat-rate pricing are genuinely attractive numbers. VentureBeat notes startup Krea's new Krea 2 Turbo is faster and allows open modification for commercial use, so NB2 Lite is not the only option in this space. Google's advantage is bundling: the model plugs directly into Workspace and existing Google AI subscriptions rather than requiring a separate vendor relationship.

Financially, the pricing is aggressive: at $0.034 per 1,000 images, the model undercuts the older NB1 model ($0.039) and slashes costs compared to standard NB2 ($0.067) and NB Pro ($0.134) tiers.

Gmail Live Voice Search Enters Beta

The third release is narrower in scope but broader in distribution. Gmail Live, announced at Google I/O last month, is now in beta testing for a subset of Android and iOS users, according to Engadget.

The feature lets users search their Gmail inbox by voice using natural language. Ask Gemini to find upcoming travel confirmations or locate an order number, and it processes the query without manual keyword entry. The Live icon appears in the Gmail search field; tapping it launches a fullscreen voice interface with a mute toggle and an exit button. Engadget notes there is currently a small processing delay between voice commands and results.

Full rollout to Google AI Pro and Ultra subscribers is expected sometime this summer. Google Workspace business customers can access a preview version. Neither Engadget nor 9to5Google, which spotted the rollout, has confirmed a specific general availability date.

This is part of a broader pattern Google is executing across its app suite. Docs Live will turn voice stream-of-consciousness into a structured document draft, pulling context from other apps, and is still pending. Keep Live, a similar feature for Google's notetaking app, has been spotted in the wild by Android Authority but has not formally launched.

The Consolidation Question

Three releases in one window signal Google's strategy across its enterprise offerings. The company is clearly trying to make Gemini the connective layer across enterprise workflows — video production, image generation, document creation, inbox management — rather than selling discrete AI tools.

Whether that consolidation actually lands depends on whether the individual products hold up at production scale. Omni Flash's API availability is a prerequisite for enterprise adoption, not a guarantee of it. NB2 Lite's pricing is competitive, but 1k resolution caps its use cases. Gmail Live is still in beta with an acknowledged processing lag.

The unresolved question is pricing transparency on Omni Flash video generation. Until Google publishes complete output costs, enterprise buyers cannot honestly compare the consolidated-stack argument against their existing multi-vendor setups.