Claude Sonnet 4

llm

Anthropic's mid-tier Claude 4 model, succeeding Claude 3.7 Sonnet with improved coding and instruction-following. Scores are extended-thinking, single attempt, no tools.

Metadata

org	Anthropic
released	2025-05-22
license	proprietary
kind	llm

Blog ↗

Benchmarks

GPQA	75.4	source ↗
AIME 2025	70.5	source ↗
SWE-bench	72.7	source ↗

Lineage

No recorded lineage — root or standalone model.

view in full graph →