Model benchmarks

Qwen3.6-35B-A3B-MLX-8bit local LLM performance

As of July 2026, Qwen3.6-35B-A3B-MLX-8bit runs at up to 69.2 tok/s for local inference (best of 4 community benchmark runs across 1 GPU).

oMLX

Model size

35B

Peak speed

69.2 tok/s

Average speed

56.2 tok/s

Min memory

n/a

Max context

7,944 tokens

Best quality

91.4

Benchmark runs

4

GPUs tested

1

Performance by hardware and tool

Every hardware/tool/quantization combination Qwen3.6-35B-A3B-MLX-8bit has been benchmarked on, ranked by peak token generation speed. Last updated July 2026.

HardwareToolQuantBest tok/sAvg tok/sMemoryContextQualityRuns
Apple M5 MaxoMLX69.2 tok/s56.2 tok/sn/a7,944 tokens77.64

Frequently asked questions

How fast is Qwen3.6-35B-A3B-MLX-8bit for local inference?
Across 4 community benchmark runs, Qwen3.6-35B-A3B-MLX-8bit reaches up to 69.2 tok/s and averages 56.2 tok/s, with the fastest results on Apple M5 Max.
How much memory does Qwen3.6-35B-A3B-MLX-8bit need?
The leanest observed configuration used about n/a of memory.
Which tools have been used to run Qwen3.6-35B-A3B-MLX-8bit?
Benchmarks were submitted using oMLX. Results are community-contributed and updated as new runs arrive.