Claude Code sends 33k tokens before reading the prompt; OpenCode sends 7k

An experiment comparing Claude Code and OpenCode found that Claude Code is inefficient with its cache strategy and token usage, using 33k tokens before reading the prompt, while OpenCode uses 7k. This matters for developers who rely on these tools for AI-powered coding. To mitigate this, consider using OpenCode or optimizing your cache strategy. The study's findings suggest a potential cost savings with OpenCode.