← Back to community benchmarks
gemma-4-26b-a4b-it
Performance
4k
tokens
623.2
PP tok/s
51.2
TG tok/s
6573
TTFT (ms)
26.2
Peak mem (GB)
Hardware
Chip
M2 Max (38c)
Memory
64 GB
GPU Cores
38
Software
oMLX
v0.3.8.dev2
macOS
macOS 26.4.1
Context
4,096
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 565.8 | 54.2 | 25.8 GB | view |
| 4k | 623.2 | 51.2 | 26.2 GB | current |
| 8k | 627.4 | 46.9 | 26.3 GB | view |
| 16k | 612.9 | 34.0 | 26.7 GB | view |
| 32k | 574.0 | 24.7 | 27.5 GB | view |
| 64k | 505.8 | 15.7 | 29.2 GB | view |
| 128k | 401.3 | 5.9 | 32.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 54.2 | 1.00× |
| 2× | 81.7 | 1.51× |
| 4× | 106.5 | 1.96× |