Anthropic has released Claude Sonnet 5, a mid-tier model with significantly improved agentic capabilities — including autonomous planning, tool use, and multi-step task completion. Priced at $2/$10 per million input/output tokens through August 31 (rising to $3/$10 after), it undercuts Opus 4.8, GPT-5.5, and Gemini 3.1 Pro while delivering performance close to Opus 4.8 on most benchmarks. On agentic coding, Sonnet 5 scores 63.2% vs. Opus 4.8's 69.2%, and it slightly outperforms Opus 4.8 on knowledge work tasks. Safety improvements include lower rates of hallucination, sycophancy, and susceptibility to prompt injection attacks. Sonnet 5 is now the default model for free and Pro Claude plans.
1.9K Impressions5 Comments