← Back to community benchmarks
gemma-4-26B-A4B-it-oQ4
Performance
32k
tokens
1,600
PP tok/s
70.7
TG tok/s
20482
TTFT (ms)
16.7
Peak mem (GB)
Hardware
Chip
M3 Ultra (60c)
Memory
256 GB
GPU Cores
60
Software
oMLX
v0.3.9.dev2
macOS
macOS 26.5
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,418 | 102.1 | 14.4 GB | view |
| 4k | 1,805 | 97.7 | 15.1 GB | view |
| 8k | 1,872 | 93.8 | 15.3 GB | view |
| 16k | 1,792 | 84.9 | 15.7 GB | view |
| 32k | 1,600 | 70.7 | 16.7 GB | current |
| 64k | 1,293 | 51.6 | 18.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 102.1 | 1.00× |
| 2× | 150.0 | 1.47× |
| 4× | 197.6 | 1.94× |
| 8× | 263.6 | 2.58× |