Model benchmarks

perplexity-ai--pplx-computer-qwen-3-6-35b-a3b-mlx-20260709 local LLM performance

As of September 2026, perplexity-ai--pplx-computer-qwen-3-6-35b-a3b-mlx-20260709 runs at up to 81.4 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).

oMLX
ShareRedditX

Model size

35B

Peak speed

81.4 tok/s

Average speed

80.9 tok/s

Avg PP

612.1 tok/s

Min memory

18.7 GB

Max context

65.536 tokens

Avg output / run

14.267 tokens

Avg runtime / run

3m 9s

Avg quality

75.6

Benchmark runs

2

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 2 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)
Overall75.575.675.8
Agent Workflow67.476.284.9
Code Generation58.359.360.2
Role Play & Narrative75.984.993.9
Research & Analysis81.982.382.8

Performance by hardware and tool

Every hardware/tool/quantization combination perplexity-ai--pplx-computer-qwen-3-6-35b-a3b-mlx-20260709 has been benchmarked on, ranked by peak token generation speed. Last updated September 2026.

HardwareToolQuantBest tok/sAvg tok/sMemoryContextQualityRuns
Apple M4 ProoMLX81.4 tok/s80.9 tok/s18.7 GB65.536 tokens75.62

Benchmark runs

All 2 perplexity-ai--pplx-computer-qwen-3-6-35b-a3b-mlx-20260709 runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.

Frequently asked questions

Is perplexity-ai--pplx-computer-qwen-3-6-35b-a3b-mlx-20260709 good for coding?
In our benchmarks, perplexity-ai--pplx-computer-qwen-3-6-35b-a3b-mlx-20260709 scores 59.3/100 for coding. It runs at about 80.9 tok/s, so if you want more speed, unsloth/Qwen3.8-27B-GGUF:IQ3_S is faster (~104.3 tok/s) and still scores well for coding (83.0/100).
Is perplexity-ai--pplx-computer-qwen-3-6-35b-a3b-mlx-20260709 good for agentic (tool-using) tasks?
In our benchmarks, perplexity-ai--pplx-computer-qwen-3-6-35b-a3b-mlx-20260709 scores 76.2/100 for agentic workflows. It runs at about 80.9 tok/s, so if you want more speed, Tiel-Coder-35B-A3B-Q4_K_S is faster (~165.0 tok/s) and still scores well for agentic workflows (89.3/100).
How fast is perplexity-ai--pplx-computer-qwen-3-6-35b-a3b-mlx-20260709 for local inference?
Across 2 community benchmark runs, perplexity-ai--pplx-computer-qwen-3-6-35b-a3b-mlx-20260709 reaches up to 81.4 tok/s and averages 80.9 tok/s, with the fastest results on Apple M4 Pro.
How much memory does perplexity-ai--pplx-computer-qwen-3-6-35b-a3b-mlx-20260709 need?
The leanest observed configuration used about 18.7 GB of memory.
Which tools have been used to run perplexity-ai--pplx-computer-qwen-3-6-35b-a3b-mlx-20260709?
Benchmarks were submitted using oMLX. Results are community-contributed and updated as new runs arrive.