← Back to community benchmarks

gemma-3-27b-it-qat

M3 Ultra (60c) · 256 GB · 4bit · 2026-04-03
Performance
1k
tokens
288.3
PP tok/s
34.8
TG tok/s
3555
TTFT (ms)
16.0
Peak mem (GB)
Hardware
Chip M3 Ultra (60c)
Memory 256 GB
GPU Cores 60
Software
oMLX v0.3.1
macOS macOS 26.3
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 288.3 34.8 16.0 GB current
4k 303.1 32.2 16.9 GB view
8k 302.3 30.5 17.2 GB view
16k 301.6 26.0 17.9 GB view
32k 294.3 20.3 19.1 GB view
64k 279.4 13.3 21.8 GB view
128k 251.2 3.3 27.7 GB view
195k 225.4 2.3 33.1 GB view
Batching Results
Batch Size TG tok/s Speedup
34.8 1.00×
51.7 1.49×
62.0 1.78×
68.6 1.97×