Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Scientific Figure Reasoning on CharXiv
Loading...
48.9
Accuracy
GPT4o
19.78
27.34
34.9
42.46
Feb 9, 2026
Accuracy
Updated 4d ago
Evaluation Results
Method
Method
Links
Accuracy
GPT4o
Access=Close-source
2026.02
48.9
InternVL3.5-38B
Parameters=38B
2026.02
44.4
SAYO-InternVL-8B
Base Model=InternVL3.5-8B
2026.02
43.2
Qwen3-VL-8B
Parameters=8B
2026.02
42.7
SAYO-Qwen-8B
Base Model=Qwen3-VL-8B
2026.02
42.5
SAYO-Qwen-4B
Base Model=Qwen3-VL-4B
2026.02
41.9
NoisyRollout-7B
Parameters=7B
2026.02
40.5
Qwen3-VL-30B-A3B
Parameters=30B-A3B
2026.02
39.8
InternVL3.5-14B
Parameters=14B
2026.02
38.9
Semantic-back-7B
Parameters=7B
2026.02
38.1
Qwen3-VL-4B
Parameters=4B
2026.02
38
OpenVLThinker-7B
Parameters=7B
2026.02
35.2
InternVL3.5-8B
Parameters=8B
2026.02
34.1
InternVL3.5-30B-A3B
Parameters=30B-A3B
2026.02
33.7
Kimi-VL-16B
Parameters=16B
2026.02
31.3
ViGoRL
2026.02
30.4
R1-Onevision-7B
Parameters=7B
2026.02
20.9
Feedback
Search any
task
Search any
task