← Back to community benchmarks
gemma-4-26b-a4b-it
Performance
8k
tokens
1,356
PP tok/s
96.2
TG tok/s
6042
TTFT (ms)
15.3
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
64 GB
GPU Cores
40
Software
oMLX
v0.3.9.dev2
macOS
macOS 26.5
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,127 | 105.8 | 14.3 GB | view |
| 4k | 1,299 | 101.1 | 14.8 GB | view |
| 8k | 1,356 | 96.2 | 15.3 GB | current |
| 16k | 1,291 | 88.0 | 15.6 GB | view |
| 32k | 1,115 | 76.6 | 16.5 GB | view |
| 64k | 948.0 | 60.6 | 18.5 GB | view |
| 128k | 718.5 | 42.1 | 22.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 105.8 | 1.00× |
| 2× | 153.3 | 1.45× |
| 4× | 199.4 | 1.88× |