Google Cloud put Nano Banana 2 and Nano Banana Pro into general availability on the Gemini Enterprise Agent Platform on May 29, 2026, moving the two image models that headlined Google I/O 2026 from preview into production for paying Vertex customers. The Google Cloud blog announcement confirmed 1K and 2K output as standard, 4K in preview, and a new video-file input mode that lets the models read a clip and generate context-aware stills from it.
Nano Banana 2 is the consumer-facing name for Gemini 3.1 Flash Image. Nano Banana Pro is Gemini 3 Pro Image. Both shipped to the Gemini app in February, but May 29 is the date enterprise teams can build on them without a preview waitlist or a hand-holding TAM call.
Try It: Wire Nano Banana Pro Into Your Creative Pipeline
The fastest path for a working creator team:
- Open the Gemini Enterprise console on Vertex AI and select the Pro model. The Vertex model ID stays gemini-3-pro-image-preview during the transition; the Nano Banana branding is cosmetic.
- Pick your output resolution. 2K is the default for hero shots and social, 1K covers batch product variants, and 4K (still preview) is for print-grade catalog work where commit-and-forget is risky.
- For the new video-input mode, upload an MP4 reference clip and prompt the model to extract a single frame, restyle a subject, or generate a poster shot from the action.
- Route the output to your asset manager or layout tool. Google's enterprise rollout post lists native integrations with Adobe Creative Cloud and WPP's Open Creative Platform.
Why It Matters
The bottleneck for AI image generation in creative teams was never raw quality. It was access on terms a procurement team would sign. The consumer Gemini app shipped Nano Banana 2 in February with the same world knowledge, subject consistency, and production-ready specs Google pitched at I/O 2026, but agencies running on Vertex had to wait through the spring for a GA SLA, billing controls, and the data-residency guarantees that legal teams require before they will let production work touch the model. May 29 closes that gap and makes Pro a default option on the same console where teams are already managing Imagen and Veo budgets.
Key Details
Nano Banana Pro lists at $0.134 per 2K image on the standard Vertex API and $0.24 for the 4K preview tier, per Google's published rate card. Batch and Flex pricing cut both numbers roughly in half for non-urgent workloads. The video-input feature is preview, not GA, so production pipelines should keep it behind a feature flag until it graduates.
Early enterprise customers named in the announcement include Shopify, Urban Outfitters, Adobe, and WPP. TestingCatalog's reporting surfaced the 4K preview detail ahead of the formal release notes and confirmed the Adobe and WPP integrations. For teams already benchmarking against Microsoft MAI-Image-2.5 and Imagen 4, Pro now joins that head-to-head as a Vertex-native option rather than a preview-only ceiling that fell out of every comparison the moment a contract was needed.
What to Do Next
If your stack is already on Vertex AI, swap Nano Banana Pro into one production batch this week and benchmark the 2K output against your incumbent on subject consistency and lead time. If you are on the consumer Gemini app, watch the Pro tier rollout there to follow the same upgrade path it took in February. For agencies negotiating enterprise contracts, the GA stamp is the leverage point: ask now for the data-residency and indemnification terms that were missing from the preview agreement.
