Model benchmarks
qwen3.8-27b-gsq3 local LLM performance
As of October 2026, qwen3.8-27b-gsq3 runs at up to 114.9 tok/s for local inference (best of 4 community benchmark runs across 1 GPU).
Model size
27B
Peak speed
114.9 tok/s
Average speed
114.4 tok/s
Avg PP
1112.0 tok/s
Min memory
n/a
Max context
131,072 tokens
Avg output / run
18,152 tokens
Avg runtime / run
2m 33s
Avg quality
84.1
Benchmark runs
4
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 4 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 | 83.4 | 84.1 | 85.0 |
| Agent Workflow | 79.6 | 84.8 | 90.0 |
| Code Generation | 72.0 | 77.9 | 82.2 |
| Role Play & Narrative | 83.3 | 87.2 | 90.0 |
| Research & Analysis | 84.3 | 86.6 | 88.1 |
Performance by hardware and tool
Every hardware/tool/quantization combination qwen3.8-27b-gsq3 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 5060 | capy-flash | — | 114.9 tok/s | 114.4 tok/s | n/a | 131,072 tokens | 84.1 | 4 |
Benchmark runs
All 4 qwen3.8-27b-gsq3 runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is qwen3.8-27b-gsq3 good for coding?
- In our benchmarks, qwen3.8-27b-gsq3 scores 77.9/100 for coding. It runs at about 114.4 tok/s, so if you want more speed, Qwen3.6-35B-A3B-Q4_K_XL is faster (~156.9 tok/s) and still scores well for coding (82.2/100).
- Is qwen3.8-27b-gsq3 good for agentic (tool-using) tasks?
- In our benchmarks, qwen3.8-27b-gsq3 scores 84.8/100 for agentic workflows. It runs at about 114.4 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 qwen3.8-27b-gsq3 for local inference?
- Across 4 community benchmark runs, qwen3.8-27b-gsq3 reaches up to 114.9 tok/s and averages 114.4 tok/s, with the fastest results on NVIDIA GeForce RTX 5060.
- Which tools have been used to run qwen3.8-27b-gsq3?
- Benchmarks were submitted using capy-flash. Results are community-contributed and updated as new runs arrive.