← Back to community benchmarks

Qwen2.5-Coder-32B-Instruct

M3 Max (30c) · 36 GB · 4bit · 2026-03-10
Performance
1k
tokens
102.5
PP tok/s
12.1
TG tok/s
9986
TTFT (ms)
17.9
Peak mem (GB)
Hardware
Chip M3 Max (30c)
Memory 36 GB
GPU Cores 30
Software
oMLX v0.2.6
macOS macOS 26.2
Context 1,024
Batching Results
Batch Size TG tok/s Speedup
1× 12.1 1.00×
2× 18.5 1.53×
4× 22.3 1.84×
8× 23.8 1.97×