Anthropic Launches Claude Sonnet 4.5, Touting Unrivaled Coding Capabilities
Anthropic has unveiled Claude Sonnet 4.5, boldly claiming it as the world's premier coding model. The update demonstrates marked improvements in complex task execution, sustaining focus for over 30 hours on multi-step programming challenges. Benchmarks reveal a 77.2% success rate on SWE-bench Verified, climbing to 82% with parallel processing—figures that ostensibly surpass competing offerings from OpenAI and Google.
The release accompanies a suite of developer tools, though claims of enhanced safety protocols were quickly undermined as researchers compromised the system within minutes. This advancement comes as the AI sector increasingly competes on specialized capabilities, with coding proficiency becoming a key battleground for enterprise adoption.
