Google Ships Budget Nano Banana 2 Lite and Omni Flash Models

The News

Google has widened its budget tier of generative media tools, releasing two models through its DeepMind division on 30 June. The first, Nano Banana 2 Lite, is an image generator tuned for speed and low cost. The second, Gemini Omni Flash, generates and edits video from conversational prompts. Both are available now in Google AI Studio, the Gemini API and the Gemini Enterprise Agent Platform.

Nano Banana 2 Lite, carrying the model identifier gemini-3.1-flash-lite-image, returns a text-to-image result in about 4 seconds and is priced at $0.034 per 1K-resolution image. Google calls it the recommended successor to the original Nano Banana (gemini-2.5-flash-image) and is folding it into consumer products, including Google Search AI Mode and the Gemini app.

Gemini Omni Flash, listed as gemini-omni-flash-preview, is the more ambitious of the pair. It accepts text, image and video inputs, lets users refine footage through plain-language instructions, and currently produces clips of up to 10 seconds. Pricing is set at $0.10 per second of output, which places a full 10-second clip at roughly one dollar.

Why It Matters

There is a well-worn playbook at work here. A flagship model captures the headlines, but a cheaper "lite" variant captures the developer base that actually ships products. By pricing image generation in fractions of a cent and video by the second, Google is aiming at the volume market: the teams running thousands of calls a day, not the few chasing state-of-the-art quality.

It echoes the shift that followed OpenAI's push into lower-cost "mini" models, which drew a wave of small teams onto paid APIs and normalised the idea that generative output should cost almost nothing per unit. The contest is moving from raw capability to unit economics. When a ten-second video costs about a dollar and an image costs roughly three cents, the constraint on adoption stops being budget and starts being imagination, and the advantage shifts to whoever controls distribution.

Indian Angle

For Indian developers, the price tag is the story. At $0.034 an image, a little under three rupees at current rates, a Bengaluru studio churning out product shots for an e-commerce catalogue can run its entire image pipeline for the cost of a few cups of chai a day. A ten-second clip at around 85 rupees turns short-form video, the currency of Indian social commerce, into a line item small enough to ignore. For the founders behind Meesho suppliers, Flipkart sellers and the country's vast Instagram-reel economy, that is a meaningful change.

It also sharpens the question hanging over India's sovereign-AI ambitions. Homegrown efforts such as Sarvam AI and Ola's Krutrim have concentrated on language rather than image and video, and Google's aggressive API pricing makes it harder for any local challenger to compete on cost in generative media. MeitY's IndiaAI Mission has poured public money into compute and model-building, yet the commercial reality is that most Indian startups will simply call a cheaper foreign API.

That revives a familiar tension for regulators. Routing millions of generative requests through US-hosted infrastructure stirs the data-residency and dependency concerns MeitY has flagged before. For now, cost will win, because the typical Indian founder optimises for runway, and at these prices Google has made the arithmetic very easy.

FAQ

When are the models available?

Both Nano Banana 2 Lite and Gemini Omni Flash went live on 30 June 2026 and can be used today through Google AI Studio, the Gemini API and the Gemini Enterprise Agent Platform. Nano Banana 2 Lite is also rolling out to consumer surfaces such as Google Search AI Mode and the Gemini app.

How much do they cost?

Nano Banana 2 Lite is priced at $0.034 per 1K-resolution image. Gemini Omni Flash costs $0.10 per second of video output, so a maximum-length 10-second clip works out to about one dollar.

How is this different from the original Nano Banana?

The original Nano Banana was the model identified as gemini-2.5-flash-image. Google now recommends Nano Banana 2 Lite (gemini-3.1-flash-lite-image) as its replacement, citing generation in roughly 4 seconds alongside stronger prompt adherence and text rendering.

What does it mean for Indian startups?

It lowers the cost floor for any product built on generated images or video, from e-commerce catalogues to advertising and gaming, while intensifying pressure on local model-builders who now compete against foreign APIs priced in cents.

This story was reported by Google DeepMind. Read the full original coverage at Google DeepMind.