Model leaderboard

Every model we’ve run through the standard Glif tests, scored out of 10 by an LLM judge. Overall needs 3 judged tests to rank; pick a capability to rank on it alone.

Prompt adherence

Every explicit element, attribute, and relation in the prompt is present and correct in the output. Penalize missing, merged, or substituted elements.

1
FLUX.2 Pro Image Generator
2 tests
8.5
2
Seedream V4.5
2 tests
8.5
3
FLUX.2 Flex Image Generator
2 tests
8.0
4
FLUX.2 Max Image Generator
2 tests
8.0
5
FLUX.2 Turbo Image Generator
2 tests
8.0
6
Krea 2 Image Generator
1 test
8.0
7
Nano Banana Pro Image Generator
2 tests
8.0
8
Seedance 2.0 Fast
1 test
8.0
9
Video (Veo 3.1 Lite)
1 test
8.0
10
GPT Image 1.5
6 tests
7.8
11
Nano Banana Image Generator
1 test
7.0
12
Image to Video (Kling V3 Pro)
1 test
7.0
13
P-Video
1 test
7.0
14
Qwen Image Generator with LoRA
6 tests
6.4
15
FLUX.2 Klein 9B Image Generator
2 tests
6.0
16
FLUX.1 Dev Image Generator
8 tests
5.8
17
FLUX.1 Schnell Image Generator
7 tests
4.8
18
Ideogram Layerize Text
1 test
1.0