Is qwen3.8-27b-splash good for coding?

In our benchmarks, qwen3.8-27b-splash scores 76.7/100 for coding. It runs at about 74.0 tok/s, so if you want more speed, Tiel-Coder-35B-A3B-MLX-oQ4e is faster (~97.4 tok/s) and still scores well for coding (80.9/100).

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

In our benchmarks, qwen3.8-27b-splash scores 78.5/100 for agentic workflows. It runs at about 74.0 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-splash for local inference?

Across 110 community benchmark runs, qwen3.8-27b-splash reaches up to 101.7 tok/s and averages 74.0 tok/s, with the fastest results on Apple M5 Max.

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

The leanest observed configuration used about 2.8 GB of memory (quantizations tested: 4bit).

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

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

Model benchmarks

qwen3.8-27b-splash local LLM performance

As of October 2026, qwen3.8-27b-splash runs at up to 101.7 tok/s for local inference (best of 110 community benchmark runs across 2 GPUs).

LM Studio4bit

ShareReddit X

Model size

27B

Peak speed

101.7 tok/s

Average speed

74.0 tok/s

Avg PP

288.1 tok/s

Min memory

2.8 GB

Max context

262,144 tokens

Avg output / run

17,451 tokens

Avg runtime / run

4m 32s

Avg quality

82.6

Benchmark runs

110

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 110 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	76.7	82.6	86.4
Agent Workflow	66.0	78.5	89.7
Code Generation	65.2	76.7	85.1
Role Play & Narrative	79.5	90.0	94.9
Research & Analysis	79.9	85.1	89.1

Performance by hardware and tool

Every hardware/tool/quantization combination qwen3.8-27b-splash 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
Apple M5 Max	LM Studio	4bit	101.7 tok/s	78.6 tok/s	2.8 GB	262,144 tokens	81.2	11
Apple M4 Max	LM Studio	4bit	79.1 tok/s	73.5 tok/s	n/a	65,536 tokens	82.7	97
Apple M5 Max	LM Studio	—	72.7 tok/s	71.9 tok/s	n/a	131,072 tokens	83.4	2

Benchmark runs

All 110 qwen3.8-27b-splash 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-splash good for coding?: In our benchmarks, qwen3.8-27b-splash scores 76.7/100 for coding. It runs at about 74.0 tok/s, so if you want more speed, Tiel-Coder-35B-A3B-MLX-oQ4e is faster (~97.4 tok/s) and still scores well for coding (80.9/100).
Is qwen3.8-27b-splash good for agentic (tool-using) tasks?: In our benchmarks, qwen3.8-27b-splash scores 78.5/100 for agentic workflows. It runs at about 74.0 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-splash for local inference?: Across 110 community benchmark runs, qwen3.8-27b-splash reaches up to 101.7 tok/s and averages 74.0 tok/s, with the fastest results on Apple M5 Max.
How much memory does qwen3.8-27b-splash need?: The leanest observed configuration used about 2.8 GB of memory (quantizations tested: 4bit).
Which tools have been used to run qwen3.8-27b-splash?: Benchmarks were submitted using LM Studio. Results are community-contributed and updated as new runs arrive.