← Back to community benchmarks
Gemma-4-E2B
Performance
16k
tokens
2,341
PP tok/s
47.7
TG tok/s
7000
TTFT (ms)
4.8
Peak mem (GB)
Hardware
Chip
M3 Pro (14c)
Memory
18 GB
GPU Cores
14
Software
oMLX
v0.5.2rc1
macOS
macOS 26.4.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 2,688 | 52.8 | 4.6 GB | view |
| 4k | 2,650 | 52.4 | 4.7 GB | view |
| 8k | 2,540 | 50.8 | 4.8 GB | view |
| 16k | 2,341 | 47.7 | 4.8 GB | current |
| 32k | 2,007 | 41.8 | 5.2 GB | view |
| 64k | 1,548 | 33.6 | 6.0 GB | view |
| 128k | 1,045 | 24.1 | 7.7 GB | view |
| 195k | 771.8 | 18.6 | 9.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 52.8 | 1.00× |
| 2× | 109.4 | 2.07× |
| 4× | 187.4 | 3.55× |
| 8× | 240.3 | 4.55× |