参与者简介公司

Anthropic

Anthropic PBC

使命 : AI systems that are safe, beneficial, and understandable.

AI生成的领域概念图，并非该机构或其设施的照片。

技术与贡献

Entreprise IA. Claude language models, Constitutional AI, focus on safety and interpretability.

部分旧有编辑内容仅有法文或英文版本。

这些早期里程碑未关联一手证据，仍需核实，不应视为已确认的承诺。

2025 · Claude Opus 4 and Sonnet 4 released
Anthropic released Claude Opus 4 and Sonnet 4 on 22 May 2025 as hybrid reasoning models, including extended thinking with tool use and availability through Claude, the API, Amazon Bedrock and Google Vertex AI.
2026 · Constitutional AI alignment improvements
RLHF + Constitutional optimization. Value learning.

组织与相关资料

每份参考资料均与其支持的论述关联。机构文献描述其自身工作，不构成独立验证。