claude-opus-5

Claude Opus 5 is an advanced model for deep reasoning, complex coding, research analysis and difficult creative work. It supports adaptive thinking and enables the thinking process by default; the level of task input can be adjusted through effort, and is suitable for long-cycle agents, complex software engineering, data research, document generation and multi-step tool orchestration. This model supports large contexts, prompt caching, and longer outputs, but max_tokens constrains both reflections and final answers, so output caps on older workflows should be re-evaluated. Prior to production deployment, emphasis should be placed on verifying latency, cost, tool call stability, and task boundaries.

Context

1000000 tokens

Max output

128000 tokens

Input

llm image

Output

llm