← Back to community benchmarks
gemma-4-26B-A4B-it-oQ8
Performance
16k
tokens
1,331
PP tok/s
57.2
TG tok/s
12310
TTFT (ms)
26.7
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.6
macOS
macOS 15.7.4
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,177 | 76.1 | 25.7 GB | view |
| 4k | 1,365 | 72.8 | 26.2 GB | view |
| 8k | 1,382 | 66.5 | 26.3 GB | view |
| 16k | 1,331 | 57.2 | 26.7 GB | current |
| 32k | 1,211 | 45.7 | 27.5 GB | view |
| 64k | 1,008 | 23.9 | 29.2 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 76.1 | 1.00× |
| 2× | 112.8 | 1.48× |
| 4× | 157.2 | 2.07× |
| 8× | 200.9 | 2.64× |