Google launches cheapest generative AI models: image and video at record prices

What happened?

On June 30, 2026, Google introduced two new generative AI models: Nano Banana 2 Lite (technical name gemini-3.1-flash-lite-image) for images and Gemini Omni Flash for video. Nano Banana 2 Lite generates an image in under 4 seconds at a price of $0.034 per image, while Gemini Omni Flash produces video at $0.10 per second. Both are available in Google AI Studio, the Gemini API, and the Gemini Enterprise Agent Platform, as announced by Alisa Fortin and Anish Nangia on the Google and Google Cloud blog.

Why is it important?

This launch dramatically lowers barriers for creators and businesses. Compared to its predecessor, Nano Banana 2 Lite is 80% faster (from 20 seconds to under 4) and roughly 50% cheaper. The price of $0.034 per image is the lowest on the market, surpassing competitors like DALL-E 3 ($0.04) and Stable Diffusion XL ($0.035). In video, Gemini Omni Flash directly competes with models like Sora ($0.15/second) and Runway Gen-3 ($0.12/second). To put it in perspective, generating 1,000 images with Nano Banana 2 Lite costs $34, compared to $40 with DALL-E 3 or $35 with Stable Diffusion XL. For video, one minute of content would cost $6 with Gemini Omni Flash, versus $9 with Sora or $7.20 with Runway Gen-3. This price gap is significant for high-volume creative teams.

Market consequences

This aggressive pricing strategy could trigger a price war in generative AI, forcing competitors like OpenAI and Stability AI to lower their rates. Historically, Google has used similar tactics in other markets, such as cloud storage or maps, to gain market share. Additionally, it democratizes access to visual creation tools, allowing small studios and freelancers to produce high-quality content at low cost. However, the maximum resolution of 1K limits its use in professional applications like large-format printing or cinema. In the advertising market, where Google Ads already integrates these models, ad production costs could drop by 60% or more, according to TheVortiq estimates. Small agencies could now compete with large studios in creative volume.

What readers should know

Nano Banana 2 Lite maintains good prompt adherence and character consistency, areas where cheap models often fail. It is integrated into Google products like AI Mode in Search, Gemini, NotebookLM, Google Photos, and Google Ads. For video, Gemini Omni Flash enables conversational editing, an innovative feature that allows scene modification via text commands, similar to what Google showed in prototypes like 'Generative AI for Video' in 2024. Developers can access both models via API with prices of $0.25 per million input tokens and $1.50 per million output tokens. This contrasts with DALL-E 3's API, which charges per image, not per token. The token structure allows more granular cost control.

"It's the first time we've seen an image generation model that combines 4-second speed with a price under 4 cents. This changes the game for content production at scale." — TheVortiq Analysis

Context and comparison

Google's previous model, the original Nano Banana (launched in 2025), took 20 seconds per image and cost $0.07. The 80% reduction in latency and 50% in price enables real-time workflows. In video, Gemini Omni Flash offers conversational editing, something Google had demonstrated in prototypes like 'VideoPoet' (2023) and 'Lumiere' (2024), but now in a public product. Compared to the evolution of generative AI, this launch recalls when Google reduced machine translation costs in 2016, making a previously expensive service affordable. In terms of impact, it could be comparable to the launch of GPT-3.5 models in 2022, which democratized generative text. Now, Google aims to do the same for image and video.

Recommendations for creators

For thumbnails, social media, product variations, and mass concept generation, Nano Banana 2 Lite is the optimal choice. A design studio generating 500 images per day would save about $1,500 monthly compared to DALL-E 3. For projects requiring 4K resolution or high print quality, opt for Nano Banana 2 ($0.07/image, 4K) or Nano Banana Pro ($0.12/image, 8K). Gemini Omni Flash is ideal for video prototypes and quick content, such as social media ads or short explainer videos. For cinematic productions, more expensive models like Runway Gen-3 Alpha ($0.15/second, 4K) may be necessary. It is recommended to test both models in Google AI Studio, which offers initial free credits.

Conclusion

Google is betting on volume and accessibility, lowering generative AI costs to capture market share. This move benefits creators but also raises questions about the sustainability of such low prices and the long-term impact on quality. The price war could lead to market consolidation, where only large players with economies of scale survive. For users, it is a golden time to experiment with generative AI without breaking the bank. However, we must watch whether Google can maintain these prices without sacrificing innovation or quality, as happened with its free language models that later became paid. For now, the democratization of visual creation is a reality.