Is Qwen3.8-27B-4bit good for coding?

In our benchmarks, Qwen3.8-27B-4bit scores 76.8/100 for coding. It runs at about 30.6 tok/s, so if you want more speed, Qwen3.8-27B-oQ4e-fp16-mtp is faster (~38.6 tok/s) and still scores well for coding (82.6/100).

Is Qwen3.8-27B-4bit good for agentic (tool-using) tasks?

In our benchmarks, Qwen3.8-27B-4bit scores 78.0/100 for agentic workflows. It runs at about 30.6 tok/s, so if you want more speed, Nex-N2.5-mini-APEX-Mini is faster (~88.2 tok/s) and still scores well for agentic workflows (91.3/100).

How fast is Qwen3.8-27B-4bit for local inference?

Across 9 community benchmark runs, Qwen3.8-27B-4bit reaches up to 31.9 tok/s and averages 30.6 tok/s, with the fastest results on Apple M5 Max.

How much memory does Qwen3.8-27B-4bit need?

The leanest observed configuration used about 17.8 GB of memory.

Which tools have been used to run Qwen3.8-27B-4bit?

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

Model benchmarks

Qwen3.8-27B-4bit local LLM performance

As of August 2026, Qwen3.8-27B-4bit runs at up to 31.9 tok/s for local inference (best of 9 community benchmark runs across 1 GPU).

oMLX

ShareReddit X

Model size

27B

Peak speed

31.9 tok/s

Average speed

30.6 tok/s

Min memory

17.8 GB

Max context

262,144 tokens

Avg runtime / run

13m 52s

Avg quality

81.9

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 9 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	73.9	81.9	86.6
Agent Workflow	61.4	78.0	88.6
Code Generation	66.4	76.8	83.0
Role Play & Narrative	77.3	89.3	95.5
Research & Analysis	78.1	83.4	87.5

Performance by hardware and tool

Every hardware/tool/quantization combination Qwen3.8-27B-4bit has been benchmarked on, ranked by peak token generation speed. Last updated August 2026.

Hardware	Tool	Quant	Best tok/s	Avg tok/s	Memory	Context	Quality	Runs
Apple M5 Max	oMLX	—	31.9 tok/s	30.6 tok/s	17.8 GB	262,144 tokens	81.9	9

Benchmark runs

All 9 Qwen3.8-27B-4bit runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.

Open run page →

Frequently asked questions

Is Qwen3.8-27B-4bit good for coding?: In our benchmarks, Qwen3.8-27B-4bit scores 76.8/100 for coding. It runs at about 30.6 tok/s, so if you want more speed, Qwen3.8-27B-oQ4e-fp16-mtp is faster (~38.6 tok/s) and still scores well for coding (82.6/100).
Is Qwen3.8-27B-4bit good for agentic (tool-using) tasks?: In our benchmarks, Qwen3.8-27B-4bit scores 78.0/100 for agentic workflows. It runs at about 30.6 tok/s, so if you want more speed, Nex-N2.5-mini-APEX-Mini is faster (~88.2 tok/s) and still scores well for agentic workflows (91.3/100).
How fast is Qwen3.8-27B-4bit for local inference?: Across 9 community benchmark runs, Qwen3.8-27B-4bit reaches up to 31.9 tok/s and averages 30.6 tok/s, with the fastest results on Apple M5 Max.
How much memory does Qwen3.8-27B-4bit need?: The leanest observed configuration used about 17.8 GB of memory.
Which tools have been used to run Qwen3.8-27B-4bit?: Benchmarks were submitted using oMLX. Results are community-contributed and updated as new runs arrive.