How fast is unsloth/Qwen3.6-35B-A3B-GGUF:UD-Q4_K_M for local inference?

Across 3 community benchmark runs, unsloth/Qwen3.6-35B-A3B-GGUF:UD-Q4_K_M reaches up to 330.1 tok/s and averages 303.4 tok/s, with the fastest results on Apple M5 Ultra.

How much memory does unsloth/Qwen3.6-35B-A3B-GGUF:UD-Q4_K_M need?

The leanest observed configuration used about 23.8 GB of memory (quantizations tested: Q4_K_M).

Which tools have been used to run unsloth/Qwen3.6-35B-A3B-GGUF:UD-Q4_K_M?

Benchmarks were submitted using splash. Results are community-contributed and updated as new runs arrive.

Model benchmarks

unsloth/Qwen3.6-35B-A3B-GGUF:UD-Q4_K_M local LLM performance

As of November 2026, unsloth/Qwen3.6-35B-A3B-GGUF:UD-Q4_K_M runs at up to 330.1 tok/s for local inference (best of 3 community benchmark runs across 2 GPUs).

splashQ4_K_M

ShareReddit X

Model size

35B

Peak speed

330.1 tok/s

Average speed

303.4 tok/s

Avg PP

6171.6 tok/s

Min memory

23.8 GB

Max context

65,536 tokens

Avg output / run

14,333 tokens

Avg runtime / run

48s

Avg quality

78.9

Benchmark runs

GPUs tested

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 3 runs. The low and high columns show how much the judge’s score varies between runs, and need at least two runs to display.

Task	P5 (low)	Avg	P95 (high)
Overall	75.9	78.9	82.7
Agent Workflow	75.7	82.6	89.6
Code Generation	67.6	72.4	77.9
Role Play & Narrative	76.9	79.2	81.8
Research & Analysis	76.6	81.6	85.0

Performance by hardware and tool

Every hardware/tool/quantization combination unsloth/Qwen3.6-35B-A3B-GGUF:UD-Q4_K_M has been benchmarked on, ranked by peak token generation speed. Last updated November 2026.

Hardware	Tool	Quant	Best tok/s	Avg tok/s	Memory	Context	Quality	Runs
Apple M5 Ultra	splash	Q4_K_M	330.1 tok/s	321.5 tok/s	24.0 GB	65,536 tokens	76.8	2
Apple M5 Max	splash	Q4_K_M	267.2 tok/s	267.2 tok/s	23.8 GB	65,536 tokens	83.2	1

Benchmark runs

All 3 unsloth/Qwen3.6-35B-A3B-GGUF:UD-Q4_K_M runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.

Open run page →

Frequently asked questions

How fast is unsloth/Qwen3.6-35B-A3B-GGUF:UD-Q4_K_M for local inference?: Across 3 community benchmark runs, unsloth/Qwen3.6-35B-A3B-GGUF:UD-Q4_K_M reaches up to 330.1 tok/s and averages 303.4 tok/s, with the fastest results on Apple M5 Ultra.
How much memory does unsloth/Qwen3.6-35B-A3B-GGUF:UD-Q4_K_M need?: The leanest observed configuration used about 23.8 GB of memory (quantizations tested: Q4_K_M).
Which tools have been used to run unsloth/Qwen3.6-35B-A3B-GGUF:UD-Q4_K_M?: Benchmarks were submitted using splash. Results are community-contributed and updated as new runs arrive.