← Back to community benchmarks
DeepSeek-R1-Distill-Llama-70B
Performance
4k
tokens
127.5
PP tok/s
8.8
TG tok/s
32165
TTFT (ms)
71.9
Peak mem (GB)
Hardware
Chip
M3 Ultra (60c)
Memory
256 GB
GPU Cores
60
Software
oMLX
v0.3.6
macOS
macOS 26.4.1
Context
4,096
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 123.6 | 9.4 | 70.7 GB | view |
| 4k | 127.5 | 8.8 | 71.9 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 9.4 | 1.00× |
| 2× | 12.9 | 1.37× |
| 4× | 16.1 | 1.71× |