Anthropic Permanently Keeps Promotional Pricing for Claude Sonnet 5
Anthropic has announced that it will permanently maintain the promotional launch pricing for its Claude Sonnet 5 model, canceling a 50% price increase originally scheduled for September 1st. The pricing will remain at $2 per million input tokens and $10 per million output tokens instead of rising to the planned $3 and $15. This decision provides long-term cost stability for developers and enterprises building complex agentic workflows that require multiple model calls. It also highlights the intensifying price competition in the LLM market, forcing providers to compete on cost-efficiency alongside raw performance. Claude Sonnet 5 features a 1 million token context window, a 128K maximum output, and scored 85.20% on the SWE-bench Verified coding benchmark. The model has shown strong performance in agentic tasks, approaching the capabilities of the flagship Opus 4.8 model.
## BACKGROUND
Claude Sonnet 5 is a highly capable AI model developed by Anthropic, designed with advanced agentic capabilities for tasks like coding, research, and browser automation. Benchmarks like SWE-bench Verified and OSWorld are used to evaluate how well AI models can resolve real-world software engineering issues and perform open-ended tasks in operating systems.