Anthropic
claude-sonnet-4-6
proprietaryMultimodal
Benchmark Average (0–100)
76.4%
Average of non-null benchmark scores across all evaluated tasks.
BENCHMARK SCORES
GPQA89.9%
Graduate-Level Google-Proof Q&A. PhD-level multiple-choice questions in chemistry, biology, and physics. Scored by accuracy.
SWE-bench Verified79.6%
SWE-bench Verified evaluates models on real-world software engineering tasks from GitHub issues.
MMMLU89.3%
Multilingual MMLU tests knowledge across many languages and subject areas.
BrowseComp74.7%
BrowseComp tests web browsing comprehension, measuring ability to find and synthesize information from websites.
HLE49.0%
Humanity's Last Exam (HLE) is a multi-modal benchmark testing frontier knowledge across mathematics, humanities, and natural sciences with 2,500 expert-level questions.
MMMU-Pro75.6%
MMMU-Pro is a more challenging version of MMMU with harder multimodal understanding tasks.
OTHER BENCHMARKS
OSWorld
72.5%
ARC-AGI v2
58.3%
MCP Atlas
61.3%
MODEL INFO
Organization
Anthropic
Context Window
200,000
Release Date
2026-02-17
Knowledge Cutoff
—
Parameters
—
License
proprietary
Input Price ($/M)
$3.0000
Output Price ($/M)
$15.0000
METADATA
Announcement Date
2026-02-17
Organization Country
US
Canonical ID
—