Dylan Di Filippo
Dylan Di FilippoAI journal

Nano Banana 2.1: generation and conversational retouch

GoogleVoice & creativity

The new version of the image model is available in the Gemini API.

General availability

Official Google logo — brand identifier, not announcement evidence
Brand logo. Google — logo officiel ; marque de son propriétaire · Image source ↗

Google’s image-generation and conversational-editing model now uses the API identifier gemini-nano-banana-2.1. The company reports improvements in visual quality, instruction following, character consistency across successive edits and text rendering, while retaining Flash-level speed and cost efficiency. These are Google’s claims, not an independent comparison performed for this article.

The documented output options include 1:4, 4:1, 1:8 and 8:1 aspect ratios at 1K, 2K and 4K resolutions. Those unusually wide or tall formats broaden the documented uses, while character consistency and readable text remain qualities to distinguish from a polished demonstration.

The October 6 release is generally available in the Gemini API. Its predecessor, gemini-3.1-flash-image, is deprecated without an announced shutdown date. Belgium appears in the supported API and AI Studio region list, but that does not establish European processing residency, access to every model or parity with all Gemini app features; account and billing conditions still apply.

← Back to this weekly edition

A journal by Dylan Di Filippo. Facts, vendor claims and editorial analysis are kept separate. Coverage is not exhaustive.