← Back to community benchmarks

gemma-4-26b-a4b-it

M5 Max (40c) · 128 GB · 4bit · 2026-04-02
Performance
64k
tokens
2,484
PP tok/s
31.7
TG tok/s
26381
TTFT (ms)
18.6
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.1
macOS macOS 26.4
Context 65,536
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 3,542 80.9 15.1 GB view
4k 3,906 79.6 15.4 GB view
8k 4,026 68.6 15.6 GB view
16k 3,844 60.9 16.0 GB view
32k 3,362 49.0 16.8 GB view
64k 2,484 31.7 18.6 GB current
Batching Results
Batch Size TG tok/s Speedup
80.9 1.00×
91.1 1.13×
153.8 1.90×
217.1 2.68×