← Back to community benchmarks
gemma-3-4b-it-qat
Performance
64k
tokens
733.2
PP tok/s
19.6
TG tok/s
89388
TTFT (ms)
5.6
Peak mem (GB)
Hardware
Chip
M3 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.1
macOS
macOS 26.4
Context
65,536
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 738.8 | 77.2 | 3.2 GB | view |
| 4k | 860.5 | 71.9 | 3.4 GB | view |
| 8k | 890.1 | 65.7 | 3.6 GB | view |
| 16k | 865.5 | 57.3 | 3.9 GB | view |
| 32k | 806.5 | 44.6 | 4.4 GB | view |
| 64k | 733.2 | 19.6 | 5.6 GB | current |
| 128k | 607.0 | 11.5 | 8.2 GB | view |
| 195k | 527.4 | 10.6 | 29.6 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 77.2 | 1.00× |
| 2× | 53.4 | 0.69× |
| 4× | 63.6 | 0.82× |
| 8× | 145.3 | 1.88× |