腾讯网

“Claude Code你就作吧，我换Codex了”

Claude Code 则更像本地终端里的高级工程助理。它贴近真实开发环境，能直接调用本地文件、命令行、MCP 和 hooks，灵活度很高。但灵活也意味着边界更模糊，一旦 MCP 配置、权限或工具链出问题，本地 Agent ...
36氪

糟糕，Claude Opus 4.7源代码级提示词曝光，底层设计全被看光

Opus 4.7如期而至，编程全面暴打GPT-5.4！CC之父亲授的性能榨干秘籍，更重磅的是，全新系统级提示词已在全网疯传。 Claude Opus 4.7，如期而至！ 距离上一代Opus 4.6，才过去两个月。迄今为止最强的Opus，直接取而代之。 在各大基准测试中，Opus 4.7在Mythos面前，略显逊色 ...

The Claude-lash is here: Opus 4.7 is burning through tokens — and some people's patience

Anthropic released a new model, Opus 4.7. Some users on X and Reddit aren't happy, calling it "dumb" and more costly on compute than previous models.

Claude Opus 4.7连夜突袭：或将抢走全球7亿打工人饭碗

视觉能力从50%跃升至接近满分！ Anthropic 正式发布 Claude Opus 4.7，核心升级落在复杂任务执行、高清视觉理解和更稳的长链路工作流上。对普通用户来说，最直接的变化是更听指令、更会看图、产出更接近成品，同时也要注意Token也会烧得更快了。 就在刚刚，Anthropic 正式发布 Claude Opus 4.7，并将它定义为当前可广泛使用的最强 Claude 模型。 性能不如 ...
腾讯网

Claude团队“暴论”：去Markdown，HTML才是AI Agent最佳交互格式

日前， Anthropic Claude Code 团队成员 Thariq Shihipar在 X 发了个小作文《Using Claude Code: The Unreasonable Effectiveness of HTML》，瞬间被 900 万 ...

Anthropic’s Claude Is Pumping Out Vulnerable Code, Cyber Experts Warn

Anthropic’s latest Claude models are introducing serious security issues into code, cyber experts say. The company is yet to officially explain why.

How Chinese coders are secretly using Claude and Gemini despite ban

Despite the unavailability of AI models like Anthropic's Claude and Google's Gemini in China, local developers are accessing ...

Anthropic releases Claude Opus 4.7, narrowly retaking lead for most powerful generally ...

Opus 4.7 utilizes an updated tokenizer that improves text processing efficiency, though it can increase the token count of certain inputs by 1.0–1.35x.

Claude Code 之父最新访谈：编程已经结束、harness 将消失，Claude Code 将 ...

Boris 回忆说，刚发布的时候 Claude Code 并没有爆火，用户不少，但增长平平。真正的拐点出现在 2025 年 5 月 Opus 4 发布的时候，指数级增长就此开始，之后每次模型升级（4.5、4.6、4.7）都会再拉一波。
Geeky Gadgets

Claude Opus 4.5 vs Gemini 3 Pro: Who Wins the Coding Tests?

What happens when two of the most advanced AI models go head-to-head in the race to redefine developer productivity? In one corner, we have Claude Opus 4.5, a powerhouse from Entropic, boasting ...

Claude Opus 4.7 Just Outperformed ChatGPT in Complex Reasoning

Discover how Claude Opus 4.7 outperforms competitors with new adaptive thinking, advanced coding capabilities, and interactive data visualization tools.

Claude Code之父Boris：AI编程时代已至，未来代码将全民化

在近期举办的AI Ascent大会上，Claude Code创始人Boris Cherny与红杉资本合伙人Lauren Reeder展开深度对话，系统阐述了AI编程工具的演进路径与产业变革趋势。这场讨论揭示了AI技术对传统软件开发模式的颠覆性影响，以及创业公司如何借助技术红利实现弯道超车。 作为AI编程领域的标志性产品，Claude ...