Gemma 4 26B (4B active)

Oct 13, 2026

Throughput

60.4 t/s gen · 818.7 tokens/s prompt processing (prefill)Prompt processing measures input prefill speed; it does not tell you the time to first token.

Quant

UD-Q4_K_M (GGUF)

KV cache

F16

System RAM

64 GB

Summary

User reports Gemma 4 26B-A4B-it at 60.4 t/s generation and 818.7 t/s prompt processing on an M4 Pro with 64 GB unified memory. Setup is dasLLAMA 0.6.4 with UD-Q4_K_M weights, f16 KV cache, and one Metal GPU, with 640 positions allocated. The figures are native PP512 and TG128 counts over five repetitions, not a single 512-input/128-output request.