Anthropic Unveils Claude Sonnet 4.5, Its Most Advanced Coding Model Yet
Key Points
- Anthropic released Claude Sonnet 4.5, billed as its most advanced coding model.
- The model claims industry‑leading results on benchmarks like SWE‑Bench Verified.
- Pricing stays at $3 per million input tokens and $15 per million output tokens.
- Claude Sonnet 4.5 can autonomously build full applications, provision databases, buy domains, and perform SOC 2 audits.
- Apple and Meta reportedly use Claude models internally; API customers include Cursor, Windsurf, and Replit.
- OpenAI’s GPT‑5 has recently outperformed Claude on some coding benchmarks.
- Anthropic introduced a Claude Agent SDK for building custom agents.
- A research preview called “Imagine with Claude” lets Max subscribers see real‑time code generation.
Anthropic announced the release of Claude Sonnet 4.5, a frontier AI model aimed at production‑ready software development. The company says the model delivers industry‑leading results on coding benchmarks such as SWE‑Bench Verified and can autonomously build full applications, provision databases, purchase domains, and even conduct SOC 2 audits. Claude Sonnet 4.5 is accessible through the Claude API and chatbot with pricing unchanged from the prior version. Anthropic also introduced a Claude Agent SDK and a research preview called “Imagine with Claude,” underscoring a rapid development cycle that positions the firm against rivals like OpenAI’s GPT‑5.
Launch Overview
On Monday Anthropic launched Claude Sonnet 4.5, a new frontier model that the company markets as its most capable AI for coding tasks. The model is offered via the Claude API and the Claude chatbot, and its pricing remains the same as Claude Sonnet 4: $3 per million input tokens and $15 per million output tokens.
Performance Claims
Anthropic states that Claude Sonnet 4.5 achieves state‑of‑the‑art performance on a range of coding benchmarks, including SWE‑Bench Verified. The firm also claims the model is the most aligned frontier AI to date, with lower rates of sycophancy, deception, and improved resistance to prompt‑injection attacks.
Researcher David Hershey explained that benchmark scores alone do not fully capture the model’s abilities. In early trials he observed the AI code autonomously for up to 30 hours, during which it built a production‑ready application, stood up database services, purchased a domain name, and performed a SOC 2 audit to verify security and compliance.
Enterprise Adoption
Anthropic’s models have become favorites among developers and enterprises. Apple and Meta reportedly use Claude models internally, and the company has sold API access to firms such as Cursor, Windsurf, and Replit for AI‑driven coding applications.
Cursor CEO Michael Truell described Claude Sonnet 4.5 as representing “state‑of‑the‑art coding performance, specifically on longer horizon tasks.” Windsurf CEO Jeff Wang called it a “new generation of coding models.”
Competitive Landscape
The release arrives amid heightened competition in the AI space. OpenAI’s GPT‑5 has recently challenged Anthropic’s dominance, outperforming Claude models on several coding benchmarks. Anthropic’s rapid development cycle—launching Claude Sonnet 4.5 less than two months after Claude Opus 4.1—demonstrates the firm’s effort to stay ahead in a market where flagship models are introduced frequently.
New Tools and SDK
Alongside the model launch, Anthropic introduced the Claude Agent SDK, the same infrastructure that powers Claude Code, enabling developers to build custom agents. The company also released a temporary research preview called “Imagine with Claude” for Max subscribers, showcasing real‑time software generation without pre‑written code.
Overall, Anthropic’s Claude Sonnet 4.5 marks a significant step forward in AI‑driven software development, combining benchmark‑leading performance with practical capabilities that aim to simplify the creation of production‑grade applications.