← Back to community benchmarks

Qwen3.5-4B-Claude-4.6-Opus-Reasoning-Distilled-v2

M3 Ultra (80c) · 512 GB · 8bit · 2026-04-15
Performance
1k
tokens
1,573
PP tok/s
105.1
TG tok/s
651
TTFT (ms)
5.3
Peak mem (GB)
Hardware
Chip M3 Ultra (80c)
Memory 512 GB
GPU Cores 80
Software
oMLX v0.3.5
macOS macOS 26.4.1
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,573 105.1 5.3 GB current
4k 2,178 101.7 6.0 GB view
8k 2,364 96.6 6.3 GB view
16k 2,277 86.7 7.1 GB view
32k 2,066 72.4 8.6 GB view
Batching Results
Batch Size TG tok/s Speedup
105.1 1.00×
177.6 1.69×
258.1 2.46×