← Back to community benchmarks
gemma-4-26B-A4B-it
Performance
32k
tokens
1,233
PP tok/s
76.5
TG tok/s
26568
TTFT (ms)
16.5
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.9
macOS
macOS 26.5
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,222 | 107.9 | 14.3 GB | view |
| 4k | 1,413 | 103.7 | 14.9 GB | view |
| 8k | 1,430 | 96.8 | 15.1 GB | view |
| 16k | 1,371 | 88.9 | 15.6 GB | view |
| 32k | 1,233 | 76.5 | 16.5 GB | current |
| 64k | 1,021 | 59.9 | 18.5 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 107.9 | 1.00× |
| 2× | 156.7 | 1.45× |
| 4× | 202.0 | 1.87× |
| 8× | 245.2 | 2.27× |