Anthropic ships Claude Opus 5.5 at $4 per million input tokens
Anthropic said the model costs 40% less than Opus 5 to run on typical workloads, and it arrives on the day a leaked identifier had pointed to.
Anthropic released Claude Opus 5.5 on Tuesday, priced at $4 per million input tokens and $20 per million output tokens. The company said the model costs 40% less than Opus 5 to run on typical workloads and generates output more than 30% faster.
Amazon said the model is available on Amazon Bedrock and on Claude Platform on AWS. Anthropic said it is available on all platforms, including Google Cloud and Microsoft Azure. Anthropic also said it is raising five-hour usage limits for subscribers on Pro, Max and Team plans.
Anthropic put the model at 66.4% on Terminal-Bench 4.0, against 52.3% for Opus 5 and 57.9% for OpenAI's GPT-6 Astra. On OSWorld 2.0, a computer-use test, it put Opus 5.5 at 81.8% against 74.0% for Opus 5. It said the model performs at the level of Claude Fable 5.1 on most tasks. The figures are its own.
Anthropic said Opus 5.5 attempted to circumvent boundaries around its testing sandbox about 85% less often than Opus 5 or Claude Mythos 5.1. It also said the model scored best of any model so far on its automated behavioural audit.
The Verge reported that Anthropic says the model carries stronger safeguards after recent rogue AI hacking incidents. It is the first Anthropic model released since the company called for pacing the frontier. The sandbox and audit claims could not be checked independently.