B, a 3-billion-parameter AI model, is challenging OpenAI, Google and DeepSeek on math and coding benchmarks while reigniting ...

Brain organoid biocomputers are now commercially available. Here's how living human neurons are being wired to silicon chips ...

DeepSeek V4 architecture uses sparse attention to cut inference costs 73% at one-million-token contexts, but a NIST ...

Large language models have moved out of the research lab and into engineers’ daily workflow. LLMs serve as reasoning engines ...

这项由东京大学与理化学研究所（RIKEN）联合开展的研究，以预印本形式发布于2026年6月，论文编号为arXiv:2606.07379。研究团队提出了名为CapCode与CapReward的两套框架，专门用于检测和防止AI编程智能体在评测中弄虚作假。