Model benchmarks

qwen3.5:27b local LLM performance

As of July 2026, qwen3.5:27b runs at up to 10.5 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).

OllamaQ4_K_M
ShareRedditX

Model size

27.8B

Peak speed

10.5 tok/s

Average speed

9.2 tok/s

Min memory

15.7 GB

Max context

8.192 tokens

Avg quality

60.6

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)
Overall58.260.662.9
Agent Workflow67.376.385.2
Code Generation4.816.828.7
Role Play & Narrative76.878.980.9
Research & Analysis65.970.474.8

Performance by hardware and tool

Every hardware/tool/quantization combination qwen3.5:27b has been benchmarked on, ranked by peak token generation speed. Last updated July 2026.

HardwareToolQuantBest tok/sAvg tok/sMemoryContextQualityRuns
Apple M4 ProOllamaQ4_K_M10.5 tok/s9.2 tok/s15.7 GB8.192 tokens60.62

Benchmark runs

All 2 qwen3.5:27b runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.

Frequently asked questions

Is qwen3.5:27b good for coding?
In our benchmarks, qwen3.5:27b scores 16.8/100 for coding. It runs at about 9.2 tok/s, so if you want more speed, /home/polaris/models/Tiel-Coder-35B-A3B-MTP-UD-Q6_K_XL.gguf is faster (~35.6 tok/s) and still scores well for coding (84.9/100).
Is qwen3.5:27b good for agentic (tool-using) tasks?
In our benchmarks, qwen3.5:27b scores 76.3/100 for agentic workflows. It runs at about 9.2 tok/s, so if you want more speed, muse-glimmer:latest is faster (~34.5 tok/s) and still scores well for agentic workflows (93.0/100).
How fast is qwen3.5:27b for local inference?
Across 2 community benchmark runs, qwen3.5:27b reaches up to 10.5 tok/s and averages 9.2 tok/s, with the fastest results on Apple M4 Pro.
How much memory does qwen3.5:27b need?
The leanest observed configuration used about 15.7 GB of memory (quantizations tested: Q4_K_M).
Which tools have been used to run qwen3.5:27b?
Benchmarks were submitted using Ollama. Results are community-contributed and updated as new runs arrive.