Claude Opus 5.5 vs GPT-6.1 Terra: benchmarks and token cost

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

Opus and Terra expose a cost-versus-score choice

Opus’s captured overall index is 58.4, compared with Terra’s 51.0. Terra’s listed input and output rates are one quarter of Opus’s. The price difference is substantial, but it should be evaluated alongside completion quality and retries. The snapshot supplies no separate coding or math index that would justify assigning either model a specialized winner label.

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.