Anthropic Releases Flagship Model Claude Opus 5.5
On September 22, Anthropic released its flagship model Claude Opus 5.5, aimed at developers and teams who want agents to handle multi-step tasks like coding and data analysis. The company says it delivers better performance and lower cost than Opus 5.
Anthropic's published pricing and the default workload cost reduction help developers estimate the migration cost for long-running agent tasks.
On September 22, Anthropic released Claude Opus 5.5, its flagship AI model for developers and teams that hand off coding, data analysis, and other multi-step tasks to agents. The company claims better performance and lower costs than Opus 5.
What's changed
In the API, a million input tokens cost $4, and output tokens cost $20. Reading from the cache costs $0,20 per million tokens. According to Anthropic's benchmarks, under default settings typical workloads will cost 40% less than with Opus 5, and response generation is more than 30% faster.
In the benchmarks Anthropic published, the new model beat Opus 5 on agentic coding, computer use, and knowledge work tasks. The company itself published most of the metrics, while the AutomationBench results for Opus 5.5 were produced and provided by Zapier.
Who's affected by the update
The release matters most to developers running long-running agent tasks, migrations, and audits of large codebases. Lower token prices and compute usage could bring down the cost of these scenarios, though the actual savings will depend on the workload and model settings.
Claude Opus 5.5 is available in Claude, Claude Code, and Claude Platform, as well as through Amazon Web Services, Google Cloud, and Microsoft Azure. Anthropic has raised the five-hour limits for Pro, Max, Team, and per-seat Enterprise plans, but didn't say what the new limits are. The announcement has no separate availability or pricing terms for users in Russia.
Sources
Source: Tproger · Программирование · tproger.ru