claude-sonnet-4-5-20250929

Claude Sonnet 4.5's date-fixed model identification is suitable for general production scenarios that require higher quality while focusing on response efficiency and cost. It can be used for complex Q&A, long document analysis, programming assistance, tool invocation, knowledge base Q&A, content generation and multi-step business processes. This version is context-aware, enabling better management of residual context in longer sessions or agent processes. Since the date suffix represents a fixed version, it is suitable for teams that require stable regression testing and repeatable output behavior; its life cycle should still be evaluated when selecting, and space should be reserved for subsequent migrations.

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

Pricing

Input: $3 / 1M tokens

Output: $15 / 1M tokens

Cached input: $0.30000000000000004 / 1M tokens

Cache write: $3.75 / 1M tokens

Endpoints

  • anthropic
  • openai

More models

  • glm-5.3
  • claude-opus-4-7

    Claude Opus 4.7 是 Opus 高能力系列中的后续版本,适合复杂分析、深度代码任务、长上下文研究与多工具协作。该代开始不再支持手动设置 budget_tokens 的传统扩展思考方式,而改用自适应思考与 effort 参数控制推理深度。它适合构建能够自行拆解任务、调用工具、检查结果并修正方案的高级智能体。迁移时需检查采样参数、结构化输出方式和最大输出长度,避免沿用旧版 API 配置造成请求失败。

  • claude-haiku-4-5-20251001

    Claude Haiku 4.5 是 Claude 家族中强调速度与高吞吐的模型,适合对延迟敏感、调用量大的生产应用。它可用于实时客服、意图识别、文本分类、信息抽取、短摘要、内容安全初筛、批量文档处理和工作流路由。相较于大型 Opus 模型,它更适合完成结构清晰、可快速验证的步骤,也可作为多模型架构的前置筛选层。它支持手动扩展思考,可在速度与复杂任务质量之间调整。

  • claude-opus-4-6-1009