← Back to community benchmarks

gemma-4-31b-it

M5 Max (40c) · 128 GB · bf16 · 2026-05-05
Performance
195k
tokens
242.3
PP tok/s
4.3
TG tok/s
825481
TTFT (ms)
88.2
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.8+jang.1
macOS macOS 26.4.1
Context 200,000
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 632.5 7.3 58.8 GB view
4k 697.3 7.3 61.5 GB view
8k 750.0 7.3 62.3 GB view
16k 647.2 7.1 63.4 GB view
32k 534.2 6.7 66.0 GB view
64k 472.2 5.9 72.0 GB view
128k 342.2 5.0 84.0 GB view
195k 242.3 4.3 88.2 GB current
Batching Results
Batch Size TG tok/s Speedup
7.3 1.00×
8.9 1.22×
17.6 2.41×
33.9 4.64×