Model benchmarks
Qwen3.8-27B-UD-Q8_K_XL local LLM performance
As of October 2026, Qwen3.8-27B-UD-Q8_K_XL runs at up to 20.7 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
27B
Peak speed
20.7 tok/s
Average speed
20.6 tok/s
Avg PP
914.0 tok/s
Min memory
41.8 GB
Max context
262,144 tokens
Avg output / run
20,281 tokens
Avg runtime / run
16m 33s
Avg quality
81.0
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.
| Task | P5 (low) | Avg | P95 (high) |
|---|---|---|---|
| Overall | 77.9 | 81.0 | 84.0 |
| Agent Workflow | 61.0 | 68.9 | 76.8 |
| Code Generation | 67.2 | 71.6 | 76.0 |
| Role Play & Narrative | 92.4 | 93.7 | 95.1 |
| Research & Analysis | 88.1 | 89.6 | 91.2 |
Performance by hardware and tool
Every hardware/tool/quantization combination Qwen3.8-27B-UD-Q8_K_XL 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 RTX A6000 | llama.cpp | Q8_K | 20.7 tok/s | 20.6 tok/s | 41.8 GB | 262,144 tokens | 81.0 | 2 |
Benchmark runs
All 2 Qwen3.8-27B-UD-Q8_K_XL runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is Qwen3.8-27B-UD-Q8_K_XL good for coding?
- In our benchmarks, Qwen3.8-27B-UD-Q8_K_XL scores 71.6/100 for coding. It runs at about 20.6 tok/s, so if you want more speed, IsValorum/Nex-N2.5-mini-APEX-I-MiniPlus-V2.1-GGUF:1 is faster (~25.2 tok/s) and still scores well for coding (84.9/100).
- Is Qwen3.8-27B-UD-Q8_K_XL good for agentic (tool-using) tasks?
- In our benchmarks, Qwen3.8-27B-UD-Q8_K_XL scores 68.9/100 for agentic workflows. It runs at about 20.6 tok/s, so if you want more speed, granite-4.2-8b-Q4_K_L is faster (~71.6 tok/s) and still scores well for agentic workflows (92.9/100).
- How fast is Qwen3.8-27B-UD-Q8_K_XL for local inference?
- Across 2 community benchmark runs, Qwen3.8-27B-UD-Q8_K_XL reaches up to 20.7 tok/s and averages 20.6 tok/s, with the fastest results on NVIDIA RTX A6000.
- How much memory does Qwen3.8-27B-UD-Q8_K_XL need?
- The leanest observed configuration used about 41.8 GB of memory (quantizations tested: Q8_K).
- Which tools have been used to run Qwen3.8-27B-UD-Q8_K_XL?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.