llamaperf

Gemma 4 26B (4B active)

on M2 8GB · Turbo-fieldfare

Tone: positiveApple hardware
Jul 30, 2026
Throughput
5-6 t/s gen
System RAM
8 GB
VRAM reported
8 GB

Summary

User reports 31-35 t/s on an M5 MacBook Pro with a custom Swift/Metal engine. Setup is an OpenAI-compatible server with streaming and tool-call support.