← Back to community benchmarks
Qwen3-Coder-Next-REAP-48B-A3B-oQ4
Performance
32k
tokens
551.8
PP tok/s
32.3
TG tok/s
59380
TTFT (ms)
30.8
Peak mem (GB)
Hardware
Chip
M4 Pro (20c)
Memory
48 GB
GPU Cores
20
Software
oMLX
v0.3.6
macOS
macOS 26.4.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 647.0 | 61.4 | 27.5 GB | view |
| 4k | 693.8 | 58.2 | 28.4 GB | view |
| 8k | 685.9 | 54.9 | 28.7 GB | view |
| 16k | 637.6 | 44.9 | 29.4 GB | view |
| 32k | 551.8 | 32.3 | 30.8 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 61.4 | 1.00× |
| 2× | 86.3 | 1.41× |
| 4× | 102.5 | 1.67× |