Opus4.6 Max
Anthropic
Parameters
—
Architecture
—
Released
—
License
—
About
No detailed description available for this model.
Benchmark Scores
| Benchmark | Score | Date | Standing |
|---|---|---|---|
|
Terminal-Bench 2.1 (Terminus-2)
coding_agent
|
84.37%
|
18.08.2026 |
#13/44
better than 72%
|
|
SWE-bench Pro
coding_agent
|
59.53%
|
18.08.2026 |
#40/77
better than 46%
|
|
NL2Repo
coding_agent
|
57.38%
|
18.08.2026 |
#16/41
better than 62%
|
|
GPQA Diamond
stem_reasoning
|
94.33%
|
18.08.2026 |
#17/80
better than 76%
|
|
Humanity's Last Exam
stem_reasoning
|
71.78%
|
18.08.2026 |
#19/71
better than 71%
|
|
LiveCodeBench v6
stem_reasoning
|
93.59%
|
18.08.2026 |
#12/49
better than 71%
|
|
IFBench (prompt loose)
instruction_following
|
62.50
|
18.08.2026 |
#14/14
better than 0%
|
|
OmniDocBench 1.5
document_understanding
|
94.53%
|
18.08.2026 |
#13/26
better than 52%
|
|
CharXiv (RQ)
document_understanding
|
66.00
|
18.08.2026 |
#33/34
better than 0%
|
|
RealWorldQA
vision_language
|
66.05%
|
18.08.2026 |
#15/23
better than 32%
|
|
AndroidWorld
general_agent
|
13.13%
|
18.08.2026 |
#11/13
better than 8%
|
|
Claw-Eval Pass^3
coding_agent
|
51.59%
|
18.08.2026 |
#11/21
better than 40%
|
|
Claw-Eval Avg
coding_agent
|
65.24%
|
18.08.2026 |
#27/40
better than 28%
|
|
QwenSWEBench
coding_agent
|
39.19%
|
18.08.2026 |
#6/9
better than 38%
|
|
CoWorkBench
general_agent
|
75.00%
|
18.08.2026 |
#7/12
better than 36%
|
|
OSWorld-Verified
general_agent
|
75.41%
|
18.08.2026 |
#11/25
better than 54%
|
|
SWE-MM
coding_agent
|
10.85%
|
18.08.2026 |
#3/4
better than 33%
|
|
MathVision
vision_language
|
37.64%
|
18.08.2026 |
#22/27
better than 15%
|
|
BabyVision
vision_language
|
12.60
|
18.08.2026 |
#24/24
better than 0%
|
|
ERQA
spatial_intelligence
|
6.80%
|
18.08.2026 |
#19/21
better than 5%
|
What is a good Terminal-Bench 2.1 (Terminus-2) score? What is a good SWE-bench Pro score? What is a good NL2Repo score?
No comments yet
Be the first to share your thoughts!