Qwen3 is seriously fast — 0.2 seconds to first token, 55 TPS, over 220% faster than the old QwQ-32B…
Related Posts
Playing “Who's the Spy” locally with an LLMIt's Chinese New Year, and everyone is talking about DeepSeek. Ruirui wants to play "Who's the Spy" but can never rou...AI TCM pulse diagnosis testManus and the AI tools I've been tinkering with latelyTrae's China version now supports DeepSeek, and it outperforms Claude3.5-Sonnet in some scenarios. QwQ-32B's performa...Learning Product with the AI Product Manager's Handbook: Reading and Knowledge NotesI came across an introduction to the second edition of the AI Product Manager’s Handbook online, found it quite thoug...MSA (Memory Sparse Attention) — an open-source way to break through the AI memory bottleneckInstead of bolting on retrieval, MSA embeds memory directly into the attention mechanism itself. That gives you end-t...
Join the discussion