Model benchmarks
unsloth/Qwen3.8-27B-UD-Q2_K_XL local LLM performance
As of November 2026, unsloth/Qwen3.8-27B-UD-Q2_K_XL runs at up to 50.6 tok/s for local inference (best of 6 community benchmark runs across 1 GPU).
Model size
27B
Peak speed
50.6 tok/s
Average speed
46.1 tok/s
Avg PP
419.1 tok/s
Min memory
13.9 GB
Max context
65,536 tokens
Avg output / run
71,287 tokens
Avg runtime / run
26m 39s
Avg quality
85.8
Benchmark runs
6
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 6 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 | 85.8 | 90.1 |
| Agent Workflow | 53.4 | 83.1 | 95.0 |
| Code Generation | 80.6 | 83.7 | 86.5 |
| Role Play & Narrative | 80.2 | 88.9 | 94.5 |
| Research & Analysis | 85.5 | 87.5 | 89.7 |
Performance by hardware and tool
Every hardware/tool/quantization combination unsloth/Qwen3.8-27B-UD-Q2_K_XL has been benchmarked on, ranked by peak token generation speed. Last updated November 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| AMD Radeon RX 7800 XT | llama.cpp | Q2_K | 50.6 tok/s | 46.1 tok/s | 13.9 GB | 65,536 tokens | 85.8 | 6 |
Benchmark runs
All 6 unsloth/Qwen3.8-27B-UD-Q2_K_XL runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is unsloth/Qwen3.8-27B-UD-Q2_K_XL good for coding?
- In our benchmarks, unsloth/Qwen3.8-27B-UD-Q2_K_XL scores 83.7/100 for coding, among the top 3 for coding on consumer hardware (≤24 GB VRAM). It runs at about 46.1 tok/s, so if you want more speed, ornith-1.5-35b-a3b-apex-mtp-i-quality is faster (~105.2 tok/s) and still scores well for coding (83.6/100).
- Is unsloth/Qwen3.8-27B-UD-Q2_K_XL good for agentic (tool-using) tasks?
- In our benchmarks, unsloth/Qwen3.8-27B-UD-Q2_K_XL scores 83.1/100 for agentic workflows. It runs at about 46.1 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 unsloth/Qwen3.8-27B-UD-Q2_K_XL for local inference?
- Across 6 community benchmark runs, unsloth/Qwen3.8-27B-UD-Q2_K_XL reaches up to 50.6 tok/s and averages 46.1 tok/s, with the fastest results on AMD Radeon RX 7800 XT.
- How much memory does unsloth/Qwen3.8-27B-UD-Q2_K_XL need?
- The leanest observed configuration used about 13.9 GB of memory (quantizations tested: Q2_K).
- Which tools have been used to run unsloth/Qwen3.8-27B-UD-Q2_K_XL?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.