Model benchmarks

huihui-qwen3.8-27b-abliterated-oq5e-mtp local LLM performance

As of September 2026, huihui-qwen3.8-27b-abliterated-oq5e-mtp runs at up to 32.2 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).

LM Studio
ShareRedditX

Model size

27B

Peak speed

32.2 tok/s

Average speed

30.9 tok/s

Avg prefill

604.6 tok/s

Min memory

18.9 GB

Max context

131,072 tokens

Avg output / run

260 tokens

Avg runtime / run

19s

Avg quality

1.6

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)
Overall0.41.62.8
Agent Workflow0.54.99.3
Code Generation0.00.00.0
Role Play & Narrative0.91.41.8
Research & Analysis0.00.00.0

Performance by hardware and tool

Every hardware/tool/quantization combination huihui-qwen3.8-27b-abliterated-oq5e-mtp has been benchmarked on, ranked by peak token generation speed. Last updated September 2026.

HardwareToolQuantBest tok/sAvg tok/sMemoryContextQualityRuns
Apple M5 MaxLM Studio32.2 tok/s30.9 tok/s18.9 GB131,072 tokens1.62

Frequently asked questions

Is huihui-qwen3.8-27b-abliterated-oq5e-mtp good for coding?
In our benchmarks, huihui-qwen3.8-27b-abliterated-oq5e-mtp scores 0.0/100 for coding. It runs at about 30.9 tok/s, so if you want more speed, K2-Horizon-MoVA-36B-A4B-MLX-4bit is faster (~48.8 tok/s) and still scores well for coding (85.3/100).
Is huihui-qwen3.8-27b-abliterated-oq5e-mtp good for agentic (tool-using) tasks?
In our benchmarks, huihui-qwen3.8-27b-abliterated-oq5e-mtp scores 4.9/100 for agentic workflows. It runs at about 30.9 tok/s, so if you want more speed, granite-4.2-8b-Q4_K_L is faster (~71.6 tok/s) and still scores well for agentic workflows (92.9/100).
How fast is huihui-qwen3.8-27b-abliterated-oq5e-mtp for local inference?
Across 2 community benchmark runs, huihui-qwen3.8-27b-abliterated-oq5e-mtp reaches up to 32.2 tok/s and averages 30.9 tok/s, with the fastest results on Apple M5 Max.
How much memory does huihui-qwen3.8-27b-abliterated-oq5e-mtp need?
The leanest observed configuration used about 18.9 GB of memory.
Which tools have been used to run huihui-qwen3.8-27b-abliterated-oq5e-mtp?
Benchmarks were submitted using LM Studio. Results are community-contributed and updated as new runs arrive.