← Back to community benchmarks

gemma-4-31b-it

M5 Max (40c) · 128 GB · 4bit · 2026-05-19
Performance
195k
tokens
265.3
PP tok/s
7.6
TG tok/s
754003
TTFT (ms)
57.8
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.9rc1
macOS macOS 26.4.1
Context 200,000
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 678.3 26.7 17.6 GB view
4k 702.9 25.8 20.2 GB view
8k 669.3 24.4 20.7 GB view
16k 580.9 21.2 21.9 GB view
32k 533.7 18.6 24.9 GB view
64k 457.0 14.7 30.8 GB view
128k 341.8 10.4 42.8 GB view
195k 265.3 7.6 57.8 GB current
Batching Results
Batch Size TG tok/s Speedup
26.7 1.00×
41.1 1.54×
50.4 1.89×
101.5 3.80×