Model benchmarks

Qwen3.6-27B-UD-MLX-4bit local LLM performance

As of July 2026, Qwen3.6-27B-UD-MLX-4bit runs at up to 20.7 tok/s for local inference (best of 7 community benchmark runs across 1 GPU).

oMLX
ShareRedditX

Model size

27B

Peak speed

20.7 tok/s

Average speed

12.1 tok/s

Min memory

n/a

Max context

65,536 tokens

Avg quality

75.2

Benchmark runs

7

GPUs tested

1

Quality by task

Average LLM-judged quality (0–100) with the run-to-run spread shown as a P5–P95 band, overall and for each benchmark task, across all 7 runs. The low and high columns show how much the judge’s score varies between runs, and need at least two runs to display.

TaskP5 (low)AvgP95 (high)
Overall63.275.281.9
Agent Workflow70.478.388.9
Code Generation63.169.678.7
Role Play & Narrative23.574.793.1
Research & Analysis65.678.485.0

Performance by hardware and tool

Every hardware/tool/quantization combination Qwen3.6-27B-UD-MLX-4bit has been benchmarked on, ranked by peak token generation speed. Last updated July 2026.

HardwareToolQuantBest tok/sAvg tok/sMemoryContextQualityRuns
Apple M5 MaxoMLX20.7 tok/s12.1 tok/sn/a65,536 tokens75.27

Frequently asked questions

Is Qwen3.6-27B-UD-MLX-4bit good for coding?
In our benchmarks, Qwen3.6-27B-UD-MLX-4bit scores 69.6/100 for coding. It runs at about 12.1 tok/s, so if you want more speed, Gemma4:E2B/QAT-MTP@131K is faster (~303.9 tok/s) and still scores well for coding (66.9/100).
Is Qwen3.6-27B-UD-MLX-4bit good for agentic (tool-using) tasks?
In our benchmarks, Qwen3.6-27B-UD-MLX-4bit scores 78.3/100 for agentic workflows. It runs at about 12.1 tok/s, so if you want more speed, Gemma4:E2B/QAT-MTP@131K is faster (~303.9 tok/s) and still scores well for agentic workflows (71.7/100).
How fast is Qwen3.6-27B-UD-MLX-4bit for local inference?
Across 7 community benchmark runs, Qwen3.6-27B-UD-MLX-4bit reaches up to 20.7 tok/s and averages 12.1 tok/s, with the fastest results on Apple M5 Max.
How much memory does Qwen3.6-27B-UD-MLX-4bit need?
The leanest observed configuration used about n/a of memory.
Which tools have been used to run Qwen3.6-27B-UD-MLX-4bit?
Benchmarks were submitted using oMLX. Results are community-contributed and updated as new runs arrive.