Event ·Official first-hand

Claude Opus 4.8 发布：自我纠错能力大幅提升

Anthropic 发布了 Claude Opus 4.8，相比 Opus 4.7 在各项能力上均有提升。最关键的改进是模型自我审查能力显著增强，能够更有效地发现自身代码中的问题，而此前版本的自审几乎无效。新模型还拥有更敏锐的判断力、更诚实的自我评估，并能更长时间独立工作。Opus 4.8 定价与上一代相同，现已可用。

Current read

对依赖 AI 编程的开发者来说，Opus 4.8 的自我纠错能力解决了代码审查的痛点——以前让模型自己 review 代码基本没用，现在可以真正信任它帮你发现 bug，建议立即升级体验。

11 sources78° AI heatLast updated 2026/5/29 01:23:35

Evidence

11 sources

Primary source歸藏(guizang.ai)

2026/5/29 01:23:35View source

Related source 1Claude

2026/5/28 17:07:33View source

Related source 2IT之家

2026/5/28 22:52:22View source

Related source 3shao__meng

2026/5/29 00:55:26View source

Related source 4Alex Albert

2026/5/28 16:59:06View source

Related source 5Mike Krieger

2026/5/28 17:10:27View source

Related source 6Jerry Liu

2026/5/28 17:17:58View source

Related source 7lmarena.ai

2026/5/28 17:30:06View source

Related source 8AI SDK

2026/5/28 17:40:42View source

Related source 9Lenny Rachitsky

2026/5/28 18:22:13View source

Related source 10Anthropic: Newsroom

2026/5/29 00:05:25View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story