← Back to community benchmarks

Qwen3.5-35B-A3B-oQ4e

M5 Max (40c) · 128 GB · 4bit · 2026-04-14
Performance
64k
tokens
1,893
PP tok/s
47.2
TG tok/s
34614
TTFT (ms)
25.7
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.3.5
macOS macOS 26.4
Context 65,536
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,971 113.3 19.9 GB view
8k 3,812 101.5 21.1 GB view
64k 1,893 47.2 25.7 GB current
Batching Results
Batch Size TG tok/s Speedup
113.3 1.00×
144.6 1.28×
236.3 2.09×