← Back to community benchmarks
gemma-4-E4B-it-ultra-uncensored-heretic4bit-mixed_4_6
Performance
16k
tokens
1,821
PP tok/s
68.8
TG tok/s
8996
TTFT (ms)
7.0
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.8
macOS
macOS 15.7.5
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 2,067 | 76.2 | 6.5 GB | view |
| 4k | 2,152 | 85.5 | 6.7 GB | view |
| 8k | 1,953 | 80.8 | 6.8 GB | view |
| 16k | 1,821 | 68.8 | 7.0 GB | current |
| 32k | 1,711 | 58.2 | 7.7 GB | view |
| 64k | 1,612 | 46.2 | 8.6 GB | view |
| 128k | 1,085 | 30.2 | 9.8 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 76.2 | 1.00× |
| 2× | 128.8 | 1.69× |
| 4× | 161.0 | 2.11× |
| 8× | 159.7 | 2.10× |