待验证50% 置信事实时间未知
MTP-LX with 4bit quantization achieves 40+ tokens per second generation speed on Mac, approximately double the speed of the MLX 6bit solution
1
来源数
50%
置信度
中期 (~90 天)
时效性
2026/7/2
首次发现
有效期至：2026/9/30
来源
Running Qwen3.6-27B Locally on Mac: 4 Solutions Benchmarked
bilibilikate人不错
相关事实
待验证Anthros MLX 6bit with Diflash achieves approximately 22 tokens per second generation speed on Mac81% 相似待验证Anthros UD Q5 GGUF quantization achieves approximately 18 tokens per second generation speed on Mac80% 相似待验证在Mac上使用MTP-LX 4bit量化运行Qwen3.6-27B，生成速度达到43.6 tok/s79% 相似待验证Ornith 35B（4bit MLX）在Mac Studio上生成速度约100 tokens/秒，比9B在Mac Mini上快近5倍77% 相似待验证在配备96GB显存的M2 Max上以1-bit精度运行时，简单任务生成速度稳定在约28-30 tokens/秒75% 相似
引用此条事实
Stable URI
https://kongchang.com/claim/55308API
curl https://kongchang.com/api/v1/knowledge/claims/55308MCP
get_claim(id=55308)