← Back to community benchmarks

HY-MT1.5-7B

M3 Ultra (80c) · 512 GB · 8bit · 2026-03-14
Performance
16k
tokens
1,205
PP tok/s
49.1
TG tok/s
13592
TTFT (ms)
10.0
Peak mem (GB)
Hardware
Chip M3 Ultra (80c)
Memory 512 GB
GPU Cores 80
Software
oMLX v0.2.10
macOS macOS 26.3.1
Context 16,384
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 1,309 70.3 8.2 GB view
4k 1,420 64.5 8.5 GB view
8k 1,352 58.4 9.0 GB view
16k 1,205 49.1 10.0 GB current
32k 980.4 37.0 12.1 GB view
64k 712.4 5.3 71.8 GB view
Batching Results
Batch Size TG tok/s Speedup
70.3 1.00×
115.3 1.64×
183.7 2.61×
225.5 3.21×