← Back to community benchmarks

Ornith-Agents-A1-3.7-35B-A3B-dare_ties_v4-oQ8e

M5 Max (40c) · 128 GB · 8bit · 2026-07-18
Performance
32k
tokens
3,366
PP tok/s
86.4
TG tok/s
9734
TTFT (ms)
38.7
Peak mem (GB)
Hardware
Chip M5 Max (40c)
Memory 128 GB
GPU Cores 40
Software
oMLX v0.5.2.dev1
macOS macOS 27.0
Context 32,768
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 3,517 100.7 35.5 GB view
4k 4,307 97.7 36.3 GB view
32k 3,366 86.4 38.7 GB current
64k 2,437 75.6 41.9 GB view
128k 1,546 59.7 47.9 GB view
Batching Results
Batch Size TG tok/s Speedup
100.7 1.00×
159.8 1.59×
242.5 2.41×