← Back to community benchmarks

Qwen3.5-35B-A3B

M1 Max (32c) · 32 GB · 4bit · 2026-03-13
Performance
1k
tokens
530.7
PP tok/s
62.2
TG tok/s
1930
TTFT (ms)
20.1
Peak mem (GB)
Hardware
Chip M1 Max (32c)
Memory 32 GB
GPU Cores 32
Software
oMLX v0.2.10
macOS macOS 26.2
Context 1,024
Batching Results
Batch Size TG tok/s Speedup
62.2 1.00×
94.2 1.51×
118.0 1.90×
133.6 2.15×