CanRunAI

开源权重本地模型

Google Gemma 3 27B IT

Near the limit of a 24GB card at 4-bit. Text estimates exclude vision overhead.

参数量
27B
推荐量化
Q4_K_M
权重估算
17.2GB
最大上下文
128K

官方模型卡

默认上下文下的量化显存

量化权重KV cache总显存
FP1654.50GB4.16GB64.11GB
Q8_029.80GB4.16GB36.94GB
Q6_K23.03GB4.16GB29.50GB
Q5_K_M19.49GB4.16GB25.60GB
Q4_K_M17.20GB4.16GB23.08GB
Q3_K_M14.30GB4.16GB19.89GB
Q2_K8.86GB4.16GB13.91GB

检查这些 GPU 能否运行该模型