Claude Opus 5.5

The first model in Anthropic's Claude 5.5 family and its first release since Dario Amodei's September 14 essay calling for the industry to pace the frontier. Anthropic says it performs at the level of Claude Fable 5.1 on most work while costing 40% less to run than Opus 5: input and output are $4 and $20 per million tokens (20% below Opus 5), cache reads $0.20 (60% below), and output streams more than 30% faster. API ID claude-opus-5-5; 1M-token context, 128K output, June 2026 knowledge cutoff. Artificial Analysis scores it 58 on Intelligence Index v4.3.2 at max effort, #1 overall, ahead of Fable 5.1 and GPT-6 Astra (53 each). Anthropic's model docs now tell developers to start with Opus 5.5 for most workloads and to reach for Fable 5.1 for demanding reasoning and long-horizon agentic work. Parameters undisclosed.

Anthropic's launch table (Opus 5.5 at max effort unless noted, against Fable 5.1 / Opus 5 / GPT-6 Astra): Terminal-Bench 4.0 66.4% at xhigh (55.8 / 52.3 / 57.9), FrontierCode v1.1 54.4%, CursorBench 4.0 57.8%, GDPval-AA v2.1 Elo 1846 (1735 / 1708 / 1542), Humanity's Last Exam with tools 67.7%, Terminal-Bench-Science 0.1 58.7% (GPT-6 Astra 64.6%), OSWorld 2.1 81.8% (partial credit), and Zapier's AutomationBench 40.0%, run without fallback models so safeguard interventions counted as failures. Results were measured with production safeguards on: when they intervened, cyber tasks fell back to Opus 4.8 and biology or frontier-LLM-development tasks to Opus 5. It is the first Opus to launch with Fable 5.1-class safeguards on cybersecurity, biology and distillation, because Anthropic judges it comparable to Claude Mythos 5.1 in biology and cybersecurity. External pre-release testing came from Frontier Design and METR. Sonnet 5.5 followed on September 28; Haiku 5.5 is due in the coming weeks.