← Back to community benchmarks
6cee5e81ee83917806bbde320786a8fb61efebee symlink
Performance
1k
tokens
1,679
PP tok/s
103.6
TG tok/s
610
TTFT (ms)
13.3
Peak mem (GB)
Hardware
Chip
M3 Ultra (60c)
Memory
96 GB
GPU Cores
60
Software
oMLX
v0.3.0
macOS
macOS 26.2
Context
1,024
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,679 | 103.6 | 13.3 GB | current |
| 4k | 1,846 | 96.2 | 13.3 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 103.6 | 1.00× |
| 2× | 103.3 | 1.00× |
| 4× | 164.3 | 1.59× |
| 8× | 233.0 | 2.25× |