Is huihui-swift15-qwen3.8-27b good for agentic (tool-using) tasks?

In our benchmarks, huihui-swift15-qwen3.8-27b scores 83.0/100 for agentic workflows. It runs at about 120.7 tok/s, so if you want more speed, ornith-1.5:35b is faster (~151.9 tok/s) and still scores well for agentic workflows (90.7/100).

How fast is huihui-swift15-qwen3.8-27b for local inference?

Across 3 community benchmark runs, huihui-swift15-qwen3.8-27b reaches up to 134.0 tok/s and averages 120.7 tok/s, with the fastest results on NVIDIA GeForce RTX 3080.

How much memory does huihui-swift15-qwen3.8-27b need?

The leanest observed configuration used about 15.5 GB of memory.

Which tools have been used to run huihui-swift15-qwen3.8-27b?

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

Model benchmarks

huihui-swift15-qwen3.8-27b local LLM performance

As of October 2026, huihui-swift15-qwen3.8-27b runs at up to 134.0 tok/s for local inference (best of 3 community benchmark runs across 1 GPU).

llama.cpp

ShareReddit X

Model size

27B

Peak speed

134.0 tok/s

Average speed

120.7 tok/s

Avg PP

1302.4 tok/s

Min memory

15.5 GB

Max context

65,536 tokens

Avg output / run

35,589 tokens

Avg runtime / run

4m 53s

Avg quality

83.6

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	80.3	83.6	88.1
Agent Workflow	76.8	83.0	90.5
Code Generation	78.3	79.7	81.5
Role Play & Narrative	81.6	87.6	94.0
Research & Analysis	80.2	84.0	86.8

Performance by hardware and tool

Every hardware/tool/quantization combination huihui-swift15-qwen3.8-27b has been benchmarked on, ranked by peak token generation speed. Last updated October 2026.

Hardware	Tool	Quant	Best tok/s	Avg tok/s	Memory	Context	Quality	Runs
NVIDIA GeForce RTX 3080	llama.cpp	—	134.0 tok/s	120.7 tok/s	15.5 GB	65,536 tokens	83.6	3

Benchmark runs

All 3 huihui-swift15-qwen3.8-27b runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.

Open run page →

Frequently asked questions

Is huihui-swift15-qwen3.8-27b good for agentic (tool-using) tasks?: In our benchmarks, huihui-swift15-qwen3.8-27b scores 83.0/100 for agentic workflows. It runs at about 120.7 tok/s, so if you want more speed, ornith-1.5:35b is faster (~151.9 tok/s) and still scores well for agentic workflows (90.7/100).
How fast is huihui-swift15-qwen3.8-27b for local inference?: Across 3 community benchmark runs, huihui-swift15-qwen3.8-27b reaches up to 134.0 tok/s and averages 120.7 tok/s, with the fastest results on NVIDIA GeForce RTX 3080.
How much memory does huihui-swift15-qwen3.8-27b need?: The leanest observed configuration used about 15.5 GB of memory.
Which tools have been used to run huihui-swift15-qwen3.8-27b?: Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.