← Back to community benchmarks

Qwen3.5-122B-MXFP4-FLOAT16

M1 Ultra (64c) · 128 GB · 4bit · 2026-03-14
User commentused mlx_lm.convert -q --q-mode mxfp4 --dtype float16
Performance
8k
tokens
500.4
PP tok/s
35.1
TG tok/s
16372
TTFT (ms)
64.8
Peak mem (GB)
Hardware
Chip M1 Ultra (64c)
Memory 128 GB
GPU Cores 64
Software
oMLX v0.2.10
macOS macOS 26.3.1
Context 8,192
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 443.9 40.5 63.5 GB view
4k 502.0 38.5 64.1 GB view
8k 500.4 35.1 64.8 GB current
16k 473.4 28.1 66.1 GB view
32k 358.0 20.5 69.0 GB view
64k 221.5 13.2 74.4 GB view
Batching Results
Batch Size TG tok/s Speedup
40.5 1.00×
57.0 1.41×
67.2 1.66×