← Back to community benchmarks
Ornith-1.0-35B
Performance
16k
tokens
1,996
PP tok/s
60.3
TG tok/s
8210
TTFT (ms)
67.0
Peak mem (GB)
Hardware
Chip
M3 Ultra (60c)
Memory
256 GB
GPU Cores
60
Software
oMLX
v0.5.1
macOS
macOS 26.5.2
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,944 | 66.0 | 65.6 GB | view |
| 4k | 2,264 | 64.3 | 66.3 GB | view |
| 8k | 2,236 | 63.7 | 66.7 GB | view |
| 16k | 1,996 | 60.3 | 67.0 GB | current |
| 32k | 1,665 | 57.2 | 67.8 GB | view |
| 64k | 1,240 | 51.2 | 68.7 GB | view |
| 128k | 823.0 | 43.1 | 70.6 GB | view |
| 195k | 605.9 | 36.5 | 73.1 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 66.0 | 1.00× |
| 2× | 68.6 | 1.04× |
| 4× | 121.5 | 1.84× |
| 8× | 240.0 | 3.64× |