First M5 Ultra benchmarks show 50 tok/s running Qwen3 27B q4 locally
Ashefromapex · reddit · 2026-09-17
A Reddit user spotted the first (unofficial) M5 Ultra benchmarks on omlx.ai: running the q4 quantization of Qwen3 27B at 8k context without MTP, it hits roughly 50 tok/s generation and 1800 tok/s prompt processing. The source's officialness is unclear but the numbers seem reasonable and promising.
More from Infra
- To dethrone Postgres, new databases must prioritize ecosystem integrations from day one — mgill25 · 2026-11-02
- ESP32-powered 4-inch touchscreen builds a TUI dashboard for monitoring a DGX inference cluster — CATLLM · 2026-11-02
- OpenRouter crosses 10 quadrillion annualized token run rate, doubling every 10.5 weeks — deedydas · 2026-11-02
- Palmer Luckey asks: can you use datacenter RAM for things prohibited by the operator's terms of service? — avaitopiper · 2026-11-02
- Disable MTP for 20-30% faster local LLM inference on limited RAM — InternalMode8159 · 2026-11-02
- Kioxia up 216% in a year but down 68% from its June 2026 peak — zephyr_z9 · 2026-11-02