← Back to community benchmarks
DeepSeek-R1-0528-Qwen3-8B
Performance
16k
tokens
165.5
PP tok/s
7.9
TG tok/s
99015
TTFT (ms)
10.8
Peak mem (GB)
Hardware
Chip
M4 (10c)
Memory
16 GB
GPU Cores
10
Software
oMLX
v0.2.13
macOS
macOS 26.3.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 216.2 | 12.3 | 8.7 GB | view |
| 4k | 208.8 | 11.3 | 9.2 GB | view |
| 8k | 195.5 | 10.3 | 9.7 GB | view |
| 16k | 165.5 | 7.9 | 10.8 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 12.3 | 1.00× |
| 2× | 23.4 | 1.90× |
| 4× | 35.3 | 2.87× |