aiaimodel.cards
← All benchmarks

Multi input (up to 5)

We haven't written a description for this benchmark yet; see the card notes below for how each maker ran it.

Your workloadRAG over docs · 12K in · 600 out · 50K requests/mo · 30% cached
  1. 1Nano Banana 2 Lite · Google
    980 Elo

“Best value”: no other model here reports a higher score at a lower cost for your workload. Scores are self-reported by each maker with different settings, so treat close calls as ties.

How each maker reported it

  • Nano Banana 2 Lite p.4–5

    Human side-by-side preference Elo (± CI omitted) from Google's model card, June 2026. Competitor names are as printed in the card ('GPT 2 Response API Low' is OpenAI's GPT Image 2 at low quality via the Responses API).

Only each model's own maker's number is shown; scores a maker reports for rival models are left out.