← Back to community benchmarks
gemma-4-26B-A4B-it-oQ6
Performance
32k
tokens
1,035
PP tok/s
25.6
TG tok/s
31664
TTFT (ms)
22.1
Peak mem (GB)
Hardware
Chip
M3 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.6
macOS
macOS 26.4.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 991.9 | 68.0 | 20.2 GB | view |
| 4k | 1,127 | 64.6 | 20.7 GB | view |
| 8k | 1,182 | 55.6 | 20.9 GB | view |
| 16k | 1,085 | 39.6 | 21.2 GB | view |
| 32k | 1,035 | 25.6 | 22.1 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 68.0 | 1.00× |
| 2× | 104.7 | 1.54× |
| 4× | 137.5 | 2.02× |
| 8× | 164.6 | 2.42× |