Claude Opus 5.5 Scores 58 on Artificial Analysis Intelligence Index

Summary

Artificial Analysis profiles Anthropic’s Claude Opus 5.5, a proprietary reasoning model released in September 2026. The model supports text and image input, produces text output, and has a 1 million-token context window. It scores 58 on the Artificial Analysis Intelligence Index, compared with a median of 25 among the comparable models shown, placing it among the leading models in the assessment. The evaluation used 260 million output tokens, versus a median of 88 million, indicating relatively high verbosity. The model is priced at $4 per million input tokens and $20 per million output tokens, with a reported average cost of $5.98 per Intelligence Index task; the page describes these prices as somewhat high relative to the relevant medians. A 95% cache discount is listed, and a blended price of $2.94 per million tokens is provided for a stated cache-hit, input, and output mix. The page identifies Claude Opus 5.5 as proprietary and says its parameter count has not been disclosed. It is available through six API providers, although pricing may vary by provider. The assessment uses Intelligence Index version 4.3.2, which combines ten evaluations covering areas such as agentic work, coding, reasoning, knowledge, and long-context performance.