← Back to context

Comment by timr

6 days ago

It's sort of amazing that Grok's image model beats Nano Banana on nearly every one of the metrics they chose to highlight.

... does it? Are you seeing something I'm not seeing? Number one is that this just doesn't appear to be true (non-lite versions beat it across the board it seems), number two is that this specifically is a low-cost bulk model and not a SOTA frontier model, of course the benchmarks are lower.

  • It’s not a universal win, but grok outperforms on 2 of the metrics, and is close on a third.

    Image editing: banana 1308, grok 1329

    Image generation: banana 1251, grok 1174 (grok loses, but is close)

    Cost: banana 3.4 cents, grok 2 cents

    Grok is a good image model, so it isn’t totally surprising, but it’s weird that they’d pick a set of metrics where they’re not the clear winner.

I was surprised with how well grok imagine performed when reviewing fast image generation. It balances quality, speed and cost really well with under $0.02 per 1K image within a few seconds.