← Back to community benchmarks
gemma-4-e4b-it
Performance
64k
tokens
3,044
PP tok/s
20.3
TG tok/s
21531
TTFT (ms)
16.0
Peak mem (GB)
Hardware
Chip
M5 Pro (20c)
Memory
64 GB
GPU Cores
20
Software
oMLX
v0.3.4
macOS
macOS 26.4
Context
65,536
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,478 | 28.0 | 14.4 GB | view |
| 4k | 3,416 | 27.4 | 14.5 GB | view |
| 8k | 3,768 | 26.9 | 14.6 GB | view |
| 16k | 3,684 | 25.7 | 14.8 GB | view |
| 32k | 3,589 | 23.6 | 15.2 GB | view |
| 64k | 3,044 | 20.3 | 16.0 GB | current |
| 128k | 2,249 | 16.4 | 17.8 GB | view |
| 195k | 1,725 | 11.8 | 19.6 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 28.0 | 1.00× |
| 2× | 38.9 | 1.39× |
| 4× | 75.8 | 2.71× |
| 8× | 143.9 | 5.14× |