← Back to community benchmarks
gemma-4-e4b-it
Performance
32k
tokens
1,145
PP tok/s
24.9
TG tok/s
28626
TTFT (ms)
9.8
Peak mem (GB)
Hardware
Chip
M1 Ultra (48c)
Memory
64 GB
GPU Cores
48
Software
oMLX
v0.3.2
macOS
macOS 26.3.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,088 | 43.5 | 8.9 GB | view |
| 4k | 1,240 | 39.5 | 9.0 GB | view |
| 8k | 1,225 | 36.8 | 9.1 GB | view |
| 16k | 1,206 | 33.3 | 9.3 GB | view |
| 32k | 1,145 | 24.9 | 9.8 GB | current |
| 64k | 1,023 | 16.6 | 10.5 GB | view |
| 128k | 837.1 | 8.7 | 12.3 GB | view |
| 195k | 707.0 | 6.8 | 14.4 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 43.5 | 1.00× |
| 2× | 61.0 | 1.40× |
| 4× | 90.5 | 2.08× |
| 8× | 106.7 | 2.45× |