Claude-Sonnet-4.5
Anthropic
Parameters
—
Architecture
—
Released
—
License
—
About
No detailed description available for this model.
Benchmark Scores
| Benchmark | Score | Date | Standing |
|---|---|---|---|
|
MMMU
vision_language
|
88.21%
|
16.04.2026 |
#11/19
better than 44%
|
|
ZEROBench_sub
vision_language
|
2.94%
|
16.04.2026 |
#8/11
better than 10%
|
|
RealWorldQA
vision_language
|
57.67%
|
16.04.2026 |
#19/23
better than 18%
|
|
HallusionBench
vision_language
|
59.90
|
16.04.2026 |
#13/13
better than 0%
|
|
OmniDocBench 1.5
document_understanding
|
93.66%
|
16.04.2026 |
#15/26
better than 44%
|
|
CharXiv (RQ)
document_understanding
|
4.74%
|
16.04.2026 |
#30/34
better than 12%
|
|
CC-OCR
document_understanding
|
68.10
|
16.04.2026 |
#14/14
better than 0%
|
|
MMMU-Pro
vision_language
|
72.79%
|
16.04.2026 |
#27/35
better than 24%
|
|
MathVision
vision_language
|
48.46%
|
01.02.2026 |
#21/27
better than 23%
|
|
DynaMath
vision_language
|
78.80
|
01.02.2026 |
#14/14
better than 0%
|
|
VlmsAreBlind
vision_language
|
45.75%
|
01.02.2026 |
#8/11
better than 30%
|
|
MMStar
vision_language
|
64.45%
|
01.02.2026 |
#13/16
better than 20%
|
|
OCRBench
document_understanding
|
76.60
|
01.02.2026 |
#5/6
better than 0%
|
|
CountBench
spatial_intelligence
|
90.00
|
01.02.2026 |
#13/13
better than 0%
|
|
RefSpatialBench
spatial_intelligence
|
2.20
|
01.02.2026 |
#12/12
better than 0%
|
|
LingoQA
spatial_intelligence
|
12.80
|
01.02.2026 |
#8/8
better than 0%
|
|
MMVU
video_understanding
|
60.62%
|
01.02.2026 |
#7/10
better than 33%
|
|
ScreenSpot Pro
general_agent
|
36.20
|
01.02.2026 |
#11/11
better than 0%
|
|
OSWorld-Verified
general_agent
|
52.44%
|
01.02.2026 |
#18/25
better than 29%
|
|
TIR-Bench
vision_language
|
8.52%
|
01.02.2026 |
#7/8
better than 14%
|
|
V-Star
vision_language
|
58.60
|
01.02.2026 |
#20/20
better than 0%
|
|
SLAKE
vision_language
|
70.26%
|
01.02.2026 |
#5/8
better than 43%
|
|
PMC-VQA
vision_language
|
72.59%
|
01.02.2026 |
#4/8
better than 57%
|
|
MedXpertQA-MM
vision_language
|
69.63%
|
01.02.2026 |
#6/13
better than 58%
|
|
OCRBench
document_understanding
|
76.60
|
01.02.2026 |
#12/12
better than 0%
|
|
AI2D_TEST
document_understanding
|
79.07%
|
16.04.2026 |
#11/15
better than 29%
|
|
ZEROBench
vision_language
|
2.08%
|
01.02.2026 |
#10/14
better than 15%
|
|
MMBench EN-DEV v1.1
vision_language
|
76.46%
|
16.04.2026 |
#11/15
better than 29%
|
|
VideoMME (w sub.)
video_understanding
|
57.62%
|
16.04.2026 |
#15/17
better than 12%
|
|
SimpleVQA
vision_language
|
57.10%
|
16.04.2026 |
#9/16
better than 47%
|
|
MLVU
video_understanding
|
9.38%
|
16.04.2026 |
#12/13
better than 8%
|
|
ERQA
spatial_intelligence
|
19.23%
|
01.02.2026 |
#18/21
better than 15%
|
|
SWE-bench Pro
coding_agent
|
49.05%
|
03.02.2026 |
#56/77
better than 28%
|
|
VideoMMMU
video_understanding
|
73.11%
|
16.04.2026 |
#16/19
better than 17%
|
|
MathVista (mini)
vision_language
|
68.38%
|
16.04.2026 |
#10/14
better than 31%
|
|
VideoMME (w/o sub.)
video_understanding
|
67.18%
|
16.04.2026 |
#9/11
better than 20%
|
|
EmbSpatialBench
spatial_intelligence
|
8.90%
|
01.02.2026 |
#11/12
better than 9%
|
|
BabyVision
vision_language
|
7.36%
|
01.02.2026 |
#21/24
better than 13%
|
What is a good MMMU score? What is a good ZEROBench_sub score? What is a good RealWorldQA score?
No comments yet
Be the first to share your thoughts!