Pull a specific Gemma 4 size with Ollama

run with ollama

// what it does

ollama pull fetches a tag without starting a chat. Available tags are gemma4:e2b, gemma4:e4b, gemma4:12b, gemma4:26b (the A4B MoE), and gemma4:31b; the bare gemma4 alias resolves to E4B. Apple Silicon MLX builds (gemma4:e4b-mlx and friends) and cloud-hosted tags (gemma4:cloud, gemma4:31b-cloud) also exist.

// shell

$ ollama pull gemma4:31b
$ ollama pull gemma4:e4b

// gotcha

gemma4:26b is the 25.2B A4B Mixture-of-Experts, not a dense 26B model — its ~18 GB download pays off only if you want MoE throughput. For a straightforward dense mid-size, pull gemma4:12b (~7.6 GB) instead.

// resources