GPT-6.1 Sol vs Claude Sonnet 5.5: benchmarks and token cost

This comparison uses GPT-6.1 Sol (Max) and Claude Sonnet 5.5 (Max, Default Fallback). It reports the captured benchmark data, with no new hands-on test.

A close index result at equal token rates

Sol scores 54.5 on the captured overall index and Sonnet scores 54.2. Both have listed rates of $2 input and $10 output per million tokens. A 0.3-point aggregate difference is a weak basis for a broad recommendation. Use representative examples, assess errors and inspect total response time to find differences that matter to the application.

How to read the table

The intelligence index is an aggregate measurement, not a percentage of correct answers in your application. Input and output prices are separate rates per million tokens. Coding and math indices are unavailable in this capture and must not be read as zero. Compare first-token latency separately from output-token throughput, and use the same model effort settings for a trial. Choose using task quality, total time and the cost of accepted results; provider features and deployment requirements need their own checks.