Benchmark result

qwen3.8:27b-mlx on Apple M5 Pro — 36.7 tok/s

Measured with Ollama 0.32.13 on August 15, 2026.

ShareReddit X

What the model built

Open full screen →

The coding scenario asks for a playable game in a single HTML file. This is exactly what the model returned, unedited.

How this run compares

1st fastest of 2 runs of this model on Apple M5 Pro · median 25.1 tok/s.

sort

136.7tok/s81.6Ollama · nvfp4this run
213.6tok/s79.2Ollama · nvfp4

Compare the nearest 5 →

qwen3.8:27b-mlx on other hardware

40.3tok/s83.8Apple M5 Max · nvfp4
29.9tok/s82.6Apple M5 Max · nvfp4
28.6tok/s81.3Apple M4 Pro · nvfp4
22.2tok/s67.5Apple M5 Max
16.3tok/s80.7Apple M4 Pro · nvfp4

Reproduce this run

toolOllama 0.32.13

modelqwen3.8:27b-mlx (nvfp4)

context32,768

clientv0.4.36+97

Set those in Ollama, then:

llm-benchmark benchmark --model "qwen3.8:27b-mlx" --tool "Ollama"

The client prompts for context length and thinking mode, and for the KV cache dtype on Unsloth Studio. Sampling is left at the model default — the values above are what the tool reported using, not overrides the client sent.