← Back to community benchmarks
gemma-4-26B-A4B-it
Performance
32k
tokens
271.9
PP tok/s
21.6
TG tok/s
120533
TTFT (ms)
15.8
Peak mem (GB)
Hardware
Chip
M4 (10c)
Memory
32 GB
GPU Cores
10
Software
oMLX
v0.3.12
macOS
macOS 26.5
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 343.6 | 29.9 | 14.2 GB | view |
| 4k | 352.1 | 29.6 | 14.8 GB | view |
| 8k | 342.1 | 28.0 | 14.8 GB | view |
| 16k | 312.2 | 24.6 | 15.0 GB | view |
| 32k | 271.9 | 21.6 | 15.8 GB | current |
| 64k | 220.1 | 16.7 | 16.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 29.9 | 1.00× |
| 2× | 44.5 | 1.49× |
| 4× | 54.0 | 1.81× |
| 8× | 63.2 | 2.11× |