In Text to Image, Nano Banana 2.1's biggest leaps are in Layout and Human Anatomy, improving on Nano Banana 2 in all 9 capabilities we measure

Our Text to Image taxonomy measures 9 capabilities and 10 use cases, each with its own leaderboard. Capabilities are the individual skills that go into making an image.

Against Nano Banana 2, Nano Banana 2.1 moves from #10 to #4 in Layout and from #6 to #3 in Human Anatomy, and makes a large gain in Text Rendering. Human Anatomy is where it sits nearest the frontier, level with GPT Image 2 (high). Lighting is the weak spot, furthest from the frontier.

➤ Layout covers flows, arrows, blocks, visual hierarchy and multi-panel compositions.

➤ Human Anatomy covers hands, faces, body proportion, dynamic poses and movement.

➤ Text Rendering covers long text, small text, symbols and artistic lettering.