Claude Opus 5.5 takes the top spot on Text Arena with 1509 points

Anthropic’s Claude Opus 5.5 just walked into Text Arena and sat down at the head of the table. The model debuted at the number one position on the competitive leaderboard with a score of 1509 plus or minus 12 points, making it the highest-ranked text-to-text model in an arena that evaluates performance across math, coding, creative writing, and other domains through community voting. The model launched on September 22, 2026, and within three days had already established itself as the benchmark other frontier models need to beat. What makes Opus 5.5 different Claude Opus 5.5 comes in at $4 per million input tokens and $20 per million output tokens. That represents roughly a 20% cost reduction compared to the previous Opus 5 model. Cache reads pricing dropped even further, landing at $0.20 per million tokens, which Anthropic says is a 20% reduction from the earlier version’s cache pricing. The model ships with a 1 million token context window and supports up to 128K output tokens, with even higher limits available through the batch API beta. On standardized benchmarks, Opus 5.5 posted a 66.4% on Terminal-Bench 4.0, an 1846 Elo rating on GDPval-AA v2.1, and 57.8% on CursorBench 4.0. I...