llm 3 View all Gemini 3 Pro sets new AI benchmarks ChatGPT vs Claude: different users, different uses How people really use ChatGPT ai research 2 View all DeepSeek-R1 reaches Nature with pure reinforcement learning Thinking Machines Lab solves LLM nondeterminism deep learning 2 View all Attention Residuals: How Kimi Rethinks Depth-Wise Information Flow in LLMs Attention Mechanisms Compared: Standard, Linear, and Flash