Anthropic has launched Claude Sonnet 5, a new mid-tier AI model that demonstrates improved agentic capabilities. This latest iteration surpasses its predecessor, Sonnet 4.6, across all benchmarks, achieving a 63.2% score on agentic coding tests, specifically SWE-bench Pro. The company also highlighted that Sonnet 5 is more cost-effective than the flagship Opus 4.8, with introductory API pricing set at $2 per million input tokens until August 31. AI
IMPACT Sets new SOTA on agentic coding benchmarks; pressures competitors to match performance and cost.
RANK_REASON Frontier-lab model release with system card [lever_c_demoted from frontier_release: ic=2 ai=1.0]
Read on Mastodon — fosstodon.org →
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →