claude-opus-4-6-1009

Claude Opus 4.6 is the proven model of the high-capacity Opus series, suitable for long-context analysis, complex coding, agent execution, visual understanding and demanding professional documentation work. It supports adaptive thinking, prompt caching and large context processing, and can be used for cross-file research, code base transformation, tool-driven process automation and complex decision-making assistance. It is more sensitive to system prompts than earlier models, so goals, constraints, callable tools, and completion conditions should be clearly described. It should be noted that traditional manual expansion thinking has been gradually abandoned, and new integration should adopt adaptive thinking and effort control.

Context
1000000 tokens
Max output
128000 tokens
Input
llm image
Output
llm

Pricing

Input: $5 / 1M tokens

Output: $25 / 1M tokens

Cached input: $0.5 / 1M tokens

Cache write: $10 / 1M tokens

Endpoints

  • anthropic
  • openai

More models

  • claude-sonnet-4-6

    Claude Sonnet 4.6 是 Sonnet 系列中面向高质量生产任务的模型,适合在性能、速度和成本之间取得平衡。它可处理长文档、代码任务、复杂工具调用、视觉输入、PDF、批量处理和智能体工作流,也适合企业知识助手、研发助手、运营自动化和客户服务升级。它支持自适应思考与大上下文,但默认是否思考取决于请求配置;部署时应明确设置输出长度、思考策略、工具权限和失败回退机制。

  • MiniMax-M2.7

    MiniMax-M2.7于2026年3月推出,是MiniMax首个深度参与自身迭代的模型。M2.7内置Agent Harness框架,能够自主构建复杂的技能集、更新自身记忆,并通过强化学习驱动迭代优化,形成“模型驱动模型进化”的闭环。一个内部版本在100余轮自我优化中将编程脚手架性能提升了30%。在软件工程基准SWE-Pro上,M2.7得分56.22%,VIBE-Pro得分55.6%,接近Opus 4.6水平。M2.7对Office三件套的复杂编辑能力显著增强,GDPval-AA ELO得分1495为开源最高。在40个复杂技能(单技能超过2000 Token)的案例中,M2.7保持了97%的技能遵循率。

  • gemini-3.5-flash-0930
  • claude-sonnet-4-6-1009