← Back to community benchmarks

Qwen3-Coder-Next

M4 Max (40c) · 128 GB · 8bit · 2026-03-10
Performance
32k
tokens
897.7
PP tok/s
42.6
TG tok/s
36503
TTFT (ms)
81.8
Peak mem (GB)
Hardware
Chip M4 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.2.7
macOS macOS 26.3.1
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 902.1 60.6 80.1 GB view
4k 1,193 59.7 80.3 GB view
8k 1,176 57.2 80.5 GB view
16k 1,059 48.3 80.9 GB view
32k 897.7 42.6 81.8 GB current
Batching Results
Batch Size TG tok/s Speedup
60.6 1.00×
101.3 1.67×
147.5 2.43×