← Back to community benchmarks
gemma-4-26B-A4B-it-oQ6
Performance
16k
tokens
1,027
PP tok/s
46.7
TG tok/s
15957
TTFT (ms)
21.5
Peak mem (GB)
Hardware
Chip
M2 Max (38c)
Memory
64 GB
GPU Cores
38
Software
oMLX
v0.3.9.dev2-memtest
macOS
macOS 26.5
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 869.3 | 53.3 | 20.3 GB | view |
| 4k | 1,024 | 52.0 | 20.7 GB | view |
| 8k | 1,077 | 50.1 | 21.1 GB | view |
| 16k | 1,027 | 46.7 | 21.5 GB | current |
| 32k | 927.8 | 41.8 | 22.5 GB | view |
| 64k | 773.6 | 34.2 | 24.5 GB | view |
| 128k | 543.0 | 26.0 | 28.7 GB | view |
| 195k | 400.8 | 18.7 | 33.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 53.3 | 1.00× |
| 2× | 81.2 | 1.52× |
| 4× | 105.8 | 1.98× |