Anthropic has launched Claude Sonnet 4.6, now the default model for Free and Pro users. The upgrade introduces a massive 1 million token context window, significantly enhancing long-document processing, multi-step reasoning, and memory retention.
Sonnet 4.6 also delivers stronger coding performance, improved instruction following, and record benchmark results including a 60.4% score on ARC-AGI-2, a key evaluation of general reasoning ability. The results indicate steady progress toward more practical, human-like AI capabilities.
Strategically, larger context windows unlock enterprise-grade use cases legal analysis, software debugging, research synthesis, and knowledge management without constant prompt resets.
With measurable gains in reasoning and reliability, Anthropic continues positioning itself as a leading force in scalable, production-ready AI systems.
