← Back to community benchmarks
gemma-4-E4B-it-PARO
Performance
64k
tokens
4,478
PP tok/s
51.3
TG tok/s
14635
TTFT (ms)
12.6
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.12
macOS
macOS 26.6
Context
65,536
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 2,014 | 99.5 | 10.6 GB | view |
| 4k | 3,869 | 100.8 | 10.6 GB | view |
| 64k | 4,478 | 51.3 | 12.6 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 99.5 | 1.00× |
| 2× | 169.1 | 1.70× |
| 4× | 251.1 | 2.52× |