Qwen
Apple's Top Mac Mini Stumbles on Qwen's Latest Model: SSD Hits AI's Ceiling
GPU sits idle 27% of the time waiting for SSD while running Qwen's MoE on 64GB M5 Mac Mini. Local AI hardware ceiling is higher than imagined.
4 articles tagged with this topic
GPU sits idle 27% of the time waiting for SSD while running Qwen's MoE on 64GB M5 Mac Mini. Local AI hardware ceiling is higher than imagined.
Chinese team releases Ling-3.1-flash: 560B total params, 25B activated per call, 1M token context. Free for two weeks, then open source.
Alibaba open-sources Qwen3-Max (2.4T params), claiming near-frontier performance. Chinese firms now have a real alternative to OpenAI or Anthropic.
DeepSeek V4-Pro: $3.48/M tokens (1/22 of GPT-5.5). Architecture, not subsidies, rewrites AI economics, making 24/7 Agents individually affordable.