Claude Sonnet 4.5

• 🏢 Anthropic's balanced Sonnet model, released 2025, for coding and agents
• 🔧 Optimized for autonomous, long-horizon agentic and tool-use workflows
• 🧠 Substantial gains in reasoning and math over prior models
• 📏 200K-token context window for long documents and codebases
• 👁️ Multimodal with vision, function calling, and web search built in
• ⚡ Extended thinking adds enhanced reasoning for complex tasks
• 🎯 Vendor reports code-edit error rate fell from 9% to 0% versus Sonnet 4
• 🔒 Operates under Anthropic's AI Safety Level 3 protections

Anthropic

Anthropic is an American artificial intelligence company headquartered in San Francisco, focused on developing large language models under the Claude family. Founded with a strong emphasis on AI safety research, the company has become one of the most…

Claude Sonnet 4.5 is the "balanced" tier of Anthropic's Claude lineup, positioned between speed-and-cost efficiency and high capability. Anthropic describes it as optimized for software coding, agentic tasks where the model runs in a loop and uses tools, and computer use, with strong showings in reasoning and mathematics. It supports a 200K-token context window, and on this catalog it carries vision, code-optimization, function-calling, web-search, and extended-thinking reasoning capabilities.

Compared with its same-family predecessor Claude Sonnet 4, Anthropic reports meaningful generational improvement: its system card notes substantial capability gains in reasoning and mathematics, plus more comprehensive safety refusals. On the company's own internal code-editing benchmark, it states the edit error rate dropped from 9% on Sonnet 4 to 0%, alongside better tool success at lower cost. Anthropic also positions it for long-horizon agentic and tool-use workflows.

Within the family, Sonnet 4.5 was succeeded by Claude Sonnet 4.6, which Anthropic positions as a full upgrade across coding, computer use, and long-context reasoning, with a 1M-token context beta and pricing matching Sonnet 4.5. For the deepest reasoning and high-autonomy work, Anthropic points users toward its Opus tier, now led by Claude Opus 4.8.

Practically, Anthropic frames Sonnet 4.5 for software engineering, agents, cybersecurity, and financial analysis where sustained reasoning and reliable tool use matter, and it remains accessible across Anthropic's API and major cloud platforms.