LLM Gateway Hub
Claude Code

For Claude Code relays, check cache pricing and rate limits first.

Coding tasks usually have long inputs, medium outputs, and repeated context. Output price alone can be misleading.

Calculate my usage
claude-code-cost.md
scenario:
  monthly_requests: 50,000
  avg_input_tokens: 3,500
  avg_output_tokens: 900
  prompt_cache_hit: 25%

watch:
  - cache price
  - model authenticity
  - rate limits
  - pricing update date

Claude Sonnet relay shortlist