← Back to community benchmarks
gemma-4-E2B-it-TurboQuant
Performance
32k
tokens
2,690
PP tok/s
154.4
TG tok/s
12184
TTFT (ms)
6.9
Peak mem (GB)
Hardware
Chip
M1 Max (32c)
Memory
32 GB
GPU Cores
32
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 | 3,208 | 179.2 | 6.3 GB | view |
| 4k | 3,242 | 185.9 | 6.3 GB | view |
| 8k | 3,174 | 148.4 | 6.4 GB | view |
| 16k | 3,005 | 171.1 | 6.5 GB | view |
| 32k | 2,690 | 154.4 | 6.9 GB | current |
| 64k | 2,195 | 118.5 | 7.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 179.2 | 1.00× |
| 2× | 101.8 | 0.57× |
| 4× | 199.1 | 1.11× |