Claude can now process entire software projects in single request, Anthropic says

Anthropic has improved its Claude Sonnet 4 AI model so it can process datasets of up to one million tokens, allowing the examination of entire software projects or dozens of research papers in one go, rather than having to slice them up.
Accessible via Anthropic’s API or Amazon Bedrock, the upgrade provides developers with the wherewithal to analyse codebases of over 75,000 lines of code, with the AI making system-wide recommendations rather than suggestions relating to individual files.
The news comes as Anthropic faces up to competition from the likes of OpenAI and Google, both of which have similar context windows, but the company notes that Claude Sonnet 4’s strength lies in its accuracy rather than capacity.
Anthropic has priced its new offering slightly higher, costing $6 p er mi ll i o nin p u tt o k e n s an d$ 22.50 per million output tokens, with the reason given that larger prompts require greater computational power.

Fast Feed