Kimi K3.1 Pro 发布了，这绝对不是小版本的迭代。因为它是在 K3 的基础上做的后训练，所以版本号是 3.1。 但实际体验下来，我没觉得它的智能水准比 GPT 6 Astra 弱多少。价格的话，是 Opus 5.5 的 50%。

Have you ever had the experience of buying a book thinking, "I'm going to start programming," only to stop after the first 20 ...

Most programming languages were written for human programmers, but that's about to change. University of Michigan professor ...

Researchers ran five models through Claude Code, OpenCode and mini-SWE-agent. Swapping harnesses changed scores no more than ...

There was a section on the entry sheet for 'languages used and the depth of experience with them.' When I had an AI draft it, ...

Now live: abliterated-model-large-flash (our version of DeepSeek-V4.1-Flash-1017). It is a cheaper alternative to our GLM-5.3 ...

Anthropic sandbox-runtime gives AI agents OS-level sandboxes without containers; Anthropic reports 84% fewer Claude Code ...

Jac combines Python and TypeScript runtimes and toolchains, while Jaseci provides a runtime for building with AI agents. Its creator says the language could reduce the complexity agents face when ...

Prompt injection test lab: build a 12-step Python harness that scores Claude, GPT, Gemini, Mistral and DeepSeek on 14 OWASP ...

An open-source optical CT platform built from LEGO bricks, a laser diode and a microcontroller lets students run real ...

TOKYO/BEIJING/AMSTERDAM >> Global firms and governments rushed on Thursday to patch critical flaws in the PHP programming ...

A critical Brotli compression bug in Node.js was patched by Google and volunteer developers, not OpenAI as headlines claimed.