Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Scientific Graphics Program Synthesis on ChartMimic 300 samples v2 (test)
Loading...
4.13
Visual Fidelity
SciTikZer-8B
3.3292
3.5371
3.745
3.9529
Apr 7, 2026
Visual Fidelity
Structural Correctness
Code Quality
Overall Preference
Updated 3mo ago
Evaluation Results
Method
Method
Links
Visual Fidelity
Structural Correctness
Code Quality
Overall Preference
SciTikZer-8B
Parameters=8B
2026.04
4.13
4.04
3.91
59
Gemini-1.5-pro
2026.04
3.82
4.02
3.78
15
GPT-4o
2026.04
3.75
3.82
3.77
11
DeTikZify-v2.5-8B
Version=v2.5, Paramete...
2026.04
3.67
3.66
3.81
13
Qwen2-VL-Instruct-32B
Parameters=32B
2026.04
3.36
2.99
3.38
2
Feedback
Search any
task
Search any
task