Anthropic released the Claude Opus 5.5 model on September 22, 2026. It immediately secured the top position on the Text Arena leaderboard, achieving a score of 1509 points with a confidence interval of plus or minus 12 points based on community voting.
The new model features a 1 million token context window and supports up to 128 000 output tokens. Users can access higher limits through the batch API beta. The platform evaluates performance across coding, mathematics, and creative writing tasks.
Pricing for the model is set at $4 per million input tokens and $20 per million output tokens, reflecting a 20% reduction compared to the previous version. Cache reads are priced at $0.20 per million tokens, also marking a 20% decrease.
In standardized testing, the model recorded 66.4% on Terminal-Bench 4.0, 57.8% on CursorBench 4.0, and an Elo rating of 1846 on GDPval-AA v2.1. Its Intelligence Index Score reached 58, with performance described as comparable to Claude Fable 5.1.
