← Back to community benchmarks
Gemma-4-26B-A4B-Hermes-VLM
Performance
32k
tokens
2,919
PP tok/s
59.4
TG tok/s
11228
TTFT (ms)
15.7
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.4
macOS
macOS 26.5
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 2,387 | 118.9 | 14.0 GB | view |
| 4k | 3,320 | 110.1 | 14.4 GB | view |
| 8k | 3,753 | 98.7 | 14.5 GB | view |
| 16k | 3,470 | 81.4 | 14.9 GB | view |
| 32k | 2,919 | 59.4 | 15.7 GB | current |
| 64k | 2,182 | 29.4 | 17.4 GB | view |
| 128k | 1,276 | 11.7 | 21.0 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 118.9 | 1.00× |
| 2× | 91.7 | 0.77× |
| 4× | 101.9 | 0.86× |
| 8× | 100.0 | 0.84× |