← Back to community benchmarks

Qwen3-8B

M5 (10c) · 24 GB · 4bit · 2026-04-24
Performance
1k
tokens
797.5
PP tok/s
26.2
TG tok/s
1284
TTFT (ms)
4.9
Peak mem (GB)
Hardware
Chip M5 (10c)
Memory 24 GB
GPU Cores 10
Software
oMLX v0.3.7
macOS macOS 26.4.1
Context 1,024
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 797.5 26.2 4.9 GB current
4k 853.6 24.0 5.4 GB view
Batching Results
Batch Size TG tok/s Speedup
26.2 1.00×
53.6 2.05×
108.9 4.16×