Claude Opus 5.5 (US)

AWS announced that Claude Opus 5.5 is now available on Amazon Bedrock via its Machine Learning Blog, confirming first-party hosting of the new Anthropic model on the AWS platform. The blog post, dated September 22, 2026, is published on the official aws.amazon.com domain and serves as the primary channel through which As the serving platform's official communication, this AWS blog post establishes the formal pathway for enterprise and developer customers to access Claude Opus 5.5 through Amazon Bedrock's managed API infrastructure. The first-party nature of the source on the AWS Machine Learning Blog provides authoritative confirmat

Anthropic introduced Claude Opus 5.5 on September 22, 2026, positioning it as the first model in the Claude 5.5 family and emphasizing capabilities in advanced agentic coding, long multi-step software tasks, and heavy knowledge workloads such as financial research, complex document analysis, and comprehensive writing. Despite the price cut, Claude Opus 5.5 remains more expensive than rival OpenAI models released the same day: GPT-6 Sol at $2/$10 per million tokens and GPT-6 Luna at $0.10/$1.50 per million tokens, though it is cheaper than OpenAI's GPT-Astra ($10 input / $50 output). CodeRabbit developer advocate Hendrik Krack noted

Anthropic has unveiled Claude Opus 5.5, its latest flagship large language model, alongside a revised pricing structure aimed at lowering costs for developers and enterprise customers. According to reporting from Tech in Asia, the release positions the model for broader commercial deployment by setting explicit, token-based rates that reduce economic barriers to high-tier AI utilization. The pricing sets Claude Opus 5.5 at US$4 per million input tokens and US$20 per million output tokens, establishing an exact 1:5 ratio between processing and generation costs. The article cites an example showing that processing 10 million input tokens and producing 1 million output tokens daily would cost about US$60 in base spending. The announcement, published on September 22, 2026, does not disclose benchmark metrics or architectural details, focusing instead on cost predictability for high-volume API integrations.

Anthropic's Claude Opus 5.5 launched on September 22, 2026 at $4 per million input tokens and $20 per million output tokens, with cache reads priced at $0.20, making it 20% cheaper than Opus 5 on base rates, 60% cheaper on cache reads, and 60% cheaper than Claude Fable 5.1 and GPT-6 Astra on list price. Anthropic says Benchmark highlights include 66.4% on Terminal-Bench 4.0, 1846 Elo on GDPval-AA v2.1, 67.7% on Humanity's Last Exam with tools, and 81.8% on OSWorld 2.0 (partial credit), though Artificial Analysis independently scored Terminal-Bench 4.0 at 59.6%, placing Opus 5.5 level with GPT-6 Astra rather than ahead. The model bre

The LLM Stats coding leaderboard ranks Claude Opus 5.5 as the top model for coding in 2026, with an arena average rating of 51.7 across seven arenas, ahead of GPT-6 Astra (49.0) and GPT-5.6 Sol (45.5). The entry lists Claude Opus 5.5 at a blended price of $7.20 per million tokens, a 1M context window, and a generation Claude Opus 5.5 leads the coding-specific benchmarks published in the leaderboard table, including a 93.9% HumanEval score and an 89.9% SWE-Bench Verified result among 116 models tested on that benchmark. The leaderboard contextualizes Opus 5.5's position relative to other frontier models such as Claude Fable 5, Claude