Running qwen 3.8 27B iq3 xxs on RTX 3060.

AI 摘要

A user successfully ran the qwen 3.8 27B iq3 xxs model on an RTX 3060, achieving 10-20 tokens per second (tps). Prompt processing for 12772 tokens took 38241.22 ms (2.99 ms per token, 333.99 tps), while evaluation for 211 tokens took 18350.47 ms (87.38 ms per token, 11.44 tps). The total time was 56591.69 ms for 12983 tokens, with a draft acceptance of 0.41250.