Назад
MODEL / GOOGLE

Gemma 3 1B

Small and efficient, but not a substitute for larger reasoning models.

Первичный источник
Всего параметров1B
Оценка памяти~1.4 GBINT4 / GGUF Q4
Контекст32KText
LicenseGemma140+ languages
01 / ЛУЧШЕ ВСЕГО
  • classification
  • summaries
  • edge
02 / ОЦЕНКА СОВМЕСТИМОСТИ
03 / ЗАПУСК ЧЕРЕЗ

These are runtime examples, not validated installation recipes. Fit estimates assume quantized files; official repositories may hold full-precision weights. MLX, GGUF and TensorRT require matching conversions or engines.

OllamaEasyollama run gemma3:1b
llama.cppMediumllama-server -m /path/to/verified-model.gguf --port 9931
LM StudioEasySearch “google/gemma-3-1b-it” → Load a Q4/INT4 build → Start Server
MLX LMMediummlx_lm.server --model google/gemma-3-1b-it
Внимание

Small and efficient, but not a substitute for larger reasoning models.