Anthropic launches Claude Sonnet 5 with enhanced agentic capabilities at 60% lower cost than flagship model

Jun 30, 2026

Anthropic has launched Claude Sonnet 5, a mid-tier AI model offering near-flagship agentic capabilities at a 60% lower cost than its Opus 4.8 model. Priced at an introductory $2 per million input tokens, Sonnet 5 narrows the performance gap on key coding and reasoning benchmarks. The release comes as Anthropic prepares for a highly anticipated IPO, having recently raised a Series H round at a $965 billion valuation, and secures a major state-wide deployment contract with California.

Claude Sonnet 5 release

▪Claude Sonnet 5 is priced at an introductory rate of $2 per million input tokens and $10 per million output tokens through August 31, 2026.
▪After August 31, 2026, Claude Sonnet 5 standard pricing will be $3 per million input tokens and $15 per million output tokens.
▪Anthropic released Claude Sonnet 5 on June 30, 2026, as a lower-priced model designed to bring enhanced agentic AI capabilities to everyday users.
▪Claude Sonnet 5 uses an updated tokenizer that can map the same input to roughly 1.0 to 1.35 times as many tokens depending on content type.

Benchmark performance gains

▪On the GDPval-AA v2 knowledge-work benchmark, Claude Sonnet 5 scored 1,618, surpassing Opus 4.8's score of 1,615 and Sonnet 4.6's score of 1,395.
▪On the Humanity's Last Exam multidisciplinary reasoning benchmark with tools, Claude Sonnet 5 scored 57.4% compared to Opus 4.8's 57.9%.
▪On a Firefox 147 exploit development evaluation, Claude Sonnet 5 scored 0.0% for working exploits but showed a 13.2% partial success rate.
▪Claude Sonnet 5 scored 63.2% on the SWE-bench Pro agentic coding benchmark, compared to 58.1% for Sonnet 4.6 and 69.2% for Opus 4.8.

Enterprise agentic AI reliability

▪Fabian Hedin of Lovable stated that Claude Sonnet 5 refuses unsafe requests cleanly and consistently.
▪Daniel Shepard of Zapier stated that Claude Sonnet 5 successfully completed a two-part automation job end-to-end that previously stalled halfway.
▪Sualeh Asif of Cursor reported that Claude Sonnet 5 allows agents to stay on plan, follow conventions, and ship clean multi-step changes.

Anthropic IPO trajectory

▪Anthropic confidentially filed its initial public offering prospectus with the United States Securities and Exchange Commission in early June 2026.
▪Anthropic raised $30 billion in February 2026 at a $380 billion post-money valuation, reporting $14 billion in annualized revenue.
▪Anthropic raised a $65 billion Series H funding round in late May 2026 at a $965 billion post-money valuation.

Government contract expansion

▪Anthropic is in ongoing discussions with the Trump administration regarding its models, including the release of Claude Sonnet 5.
▪California Governor Gavin Newsom announced a partnership providing Claude to all California state agencies at a 50% discount with free workforce training.

AI market competition

▪OpenAI launched its GPT-5.6 Sol model in preview, which is designed to allow users to split work across subagents for autonomous tasks.
▪Z.ai released GLM-5.2, an open-weight model built for long-horizon tasks with a 1-million-token context window.
▪Google launched Gemini 3.5 Flash in May 2026, pitching it as an agentic tool that plans, builds, and iterates with minimal human input.