← Back to community benchmarks
gemma-4-26B-A4B-it-oQ8
Performance
8k
tokens
2,748
PP tok/s
71.5
TG tok/s
2981
TTFT (ms)
26.5
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.6
macOS
macOS 26.5
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,986 | 83.9 | 25.7 GB | view |
| 4k | 2,673 | 79.0 | 26.3 GB | view |
| 8k | 2,748 | 71.5 | 26.5 GB | current |
| 16k | 2,693 | 59.6 | 26.9 GB | view |
| 32k | 2,523 | 40.8 | 27.6 GB | view |
| 64k | 2,202 | 27.1 | 29.2 GB | view |
| 128k | 1,684 | 7.0 | 32.7 GB | view |
| 195k | 1,239 | 5.4 | 36.3 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 83.9 | 1.00× |
| 2× | 129.2 | 1.54× |
| 4× | 179.1 | 2.13× |
| 8× | 347.2 | 4.14× |