LocalForge
Terug
MODEL / GOOGLE

Gemma 3 27B

Google's 24 GB fit example is for the INT4 variant and excludes some runtime overhead.

Primaire bron
Totale parameters27B
Geschat geheugen~20.5 GBINT4
Context128KText + Image
LicenseGemma140+ languages
01 / GESCHIKT VOOR
  • vision
  • knowledge work
  • quality chat
02 / GESCHATTE FIT
Apple silicon · 32 GBTightNVIDIA · 24 GBTightApple silicon · 64 GBComfortableWorkstation · 48 GBComfortableDatacenter · 80 GBComfortable
03 / DRAAI MET
OllamaEasyollama run gemma3:27b
llama.cppMediumllama-server -m /path/to/verified-model.gguf --port 8080
LM StudioEasySearch “google/gemma-3-27b-it” → Load a Q4/INT4 build → Start Server
MLX LMMediummlx_lm.chat --model google/gemma-3-27b-it
Let op

Google's 24 GB fit example is for the INT4 variant and excludes some runtime overhead.