Model leaderboard

Every model runs the same Glif tests. Scores come from 2,042 blind votes by 10 Glif team members, plus 1,279 AI judge scores, each counting as half a vote. Each kind of task has its own board.

Score out of 100: how often people would pick this model's output over an average model's on the same tests. 50 is average.

How we score

Every model runs the same Glif tests. Team members see two outputs side by side without the model names and pick the one that better fits the test, or call it a tie. An AI judge also scores every output from 0 to 10.

A Bradley-Terry fit turns all those head-to-head results into one rating per model. Ties count as half a win, and each AI judge score counts as half a vote.

The score is the predicted chance people prefer this model's output to the average model's on the same board, so 50 is average. A board is one kind of task, like text-to-image. Scores only compare models on the same board.

The 90% range comes from resampling the votes and refitting many times. Overlapping ranges are close to a tie. A model ranks once it has 20 votes' worth of ratings on a board. Until then it shows no score, only how close it is.

Bars show each score’s 90% range. Overlapping bars are close to a tie.

Hide vote counts

309 team votes on 6 tests. A model ranks once it has 20 votes’ worth of ratings here.

  1. 1

    GPT Image 2.5Glif default

    90% range 77–91
    Wins 43 / Losses 6 / Ties 9 · 58 team votes · 6 AI judge scores

    84 out of 100, 90% range 77 to 91

  2. 2

    FLUX 3 Image

    90% range 63–90
    Wins 17 / Losses 5 / Ties 0 · 22 team votes · 4 AI judge scores

    78 out of 100, 90% range 63 to 90

  3. 3

    Gemini 3.1 FlashGlif default

    90% range 64–84
    Wins 22 / Losses 4 / Ties 4 · 30 team votes · 6 AI judge scores

    75 out of 100, 90% range 64 to 84

  4. 4

    FLUX.2 Flex

    90% range 50–69
    Wins 25 / Losses 15 / Ties 7 · 47 team votes · 6 AI judge scores

    59 out of 100, 90% range 50 to 69

  5. 5

    Seedream 5.0 Pro

    90% range 34–57
    Wins 18 / Losses 22 / Ties 3 · 43 team votes · 5 AI judge scores

    44 out of 100, 90% range 34 to 57

  6. 6

    Nano Banana Pro

    90% range 33–53
    Wins 15 / Losses 22 / Ties 7 · 44 team votes · 5 AI judge scores

    44 out of 100, 90% range 33 to 53

  7. 7

    Phota Image Editor

    90% range 31–55
    Wins 10 / Losses 10 / Ties 5 · 25 team votes · 5 AI judge scores

    43 out of 100, 90% range 31 to 55

  8. 8

    Ideogram V4.5Glif default

    90% range 28–53
    Wins 16 / Losses 25 / Ties 13 · 54 team votes · 5 AI judge scores

    40 out of 100, 90% range 28 to 53

  9. 9

    Muse ImageGlif default

    90% range 25–45
    Wins 10 / Losses 23 / Ties 9 · 42 team votes · 4 AI judge scores

    33 out of 100, 90% range 25 to 45

  10. 10

    Seedream V4.5

    90% range 20–41
    Wins 4 / Losses 12 / Ties 8 · 24 team votes · 5 AI judge scores

    30 out of 100, 90% range 20 to 41

  11. 11

    Grok Imagine Image

    90% range 16–31
    Wins 1 / Losses 15 / Ties 11 · 27 team votes · 6 AI judge scores

    22 out of 100, 90% range 16 to 31

  12. 12

    FLUX.2 Pro

    90% range 15–26
    Wins 1 / Losses 15 / Ties 10 · 26 team votes · 6 AI judge scores

    20 out of 100, 90% range 15 to 26

  13. 13

    Z-Image Generator

    90% range 12–27
    Wins 3 / Losses 16 / Ties 7 · 26 team votes · 5 AI judge scores

    18 out of 100, 90% range 12 to 27

  14. 14

    Seedream 5.0 Lite

    90% range 10–24
    Wins 1 / Losses 17 / Ties 7 · 25 team votes · 5 AI judge scores

    17 out of 100, 90% range 10 to 24

Not ranked yet (2)