tokenstat
tokenstat

Compare

Claude Sonnet 5 Thinking vs GPT-5.5

List rates and scores as of . Lower $/MTok is cheaper. Higher scores rank better. Input edge: GPT-5.5 Output edge: GPT-5.5

Model A

Claude Sonnet 5 Thinking

Anthropic
anthropic/claude-sonnet-5
reasoningtools
  • List input$1
  • List output$5
  • Context1M
  • Offers62
  • Arena code1537
  • Arena agent5.97

On plans: Amazon Q Developer Pro, Claude Enterprise, Claude Max $100, Claude Max $200

Model B

GPT-5.5

OpenAI
openai/gpt-5.5
reasoningtools
  • List input$0.1875
  • List output$1.13
  • Context1.1M
  • Offers57
  • Arena codeN/A
  • Arena agent5.03

On plans: Cursor Business, Cursor Pro, Cursor Ultra, Copilot Business

Score sources on /benchmarks.