Coverage (30d)
1vs4
This Week
1vs1
Evidence
1 articlesRelationships
1Timeline
Claude Sonnet 52026-06-30
Anthropic released Claude Sonnet 5, beating Opus 4.8 on GDPval-AA v2
Claude Sonnet 4.62026-04-16
Outperformed GPT-4o in real-world tests on multi-file development tasks
Claude Sonnet 4.62026-04-11
Independent benchmarks validate Claude Sonnet 4.6 as a top-tier model for complex reasoning and coding tasks.
Claude Sonnet 4.62026-04-06
Showed only 3.7% self-preservation bias in a study testing AI deception, the lowest among prominent models tested.
Claude Sonnet 4.62026-03-26
Used in prompt compression study analyzing 358 successful runs from 1,199 real orchestration instructions
Claude Sonnet 4.62026-03-20
Anthropic released Claude Sonnet 4.6 with native chain-of-thought reasoning mode for complex coding tasks
Claude Sonnet 4.62026-03-17
Service disruption with elevated error rates reported on status page
Ecosystem
Claude Sonnet 5
competes withClaude Opus 4.61 src
competes withClaude Sonnet 4.61 src
Claude Sonnet 4.6
deploysChain-of-Thought Prompting1 src
deploysConstitutional AI1 src
Benchmarks
mmlu pro
Claude Sonnet 5—
Claude Sonnet 4.685
arena elo
Claude Sonnet 5—
Claude Sonnet 4.61470
osworld-verified
Claude Sonnet 5—
Claude Sonnet 4.672.1
swe bench verified
Claude Sonnet 5—
Claude Sonnet 4.679.6