Model benchmarks
JonathanColetti-Qwen3.8-27B-Uncensored-GGUF-Q5_K_M local LLM performance
As of October 2026, JonathanColetti-Qwen3.8-27B-Uncensored-GGUF-Q5_K_M runs at up to 50.8 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
27B
Peak speed
50.8 tok/s
Average speed
40.4 tok/s
Avg PP
478.7 tok/s
Min memory
31.0 GB
Max context
262,144 tokens
Avg output / run
79,786 tokens
Avg runtime / run
34m 37s
Avg quality
89.1
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 | 87.8 | 89.1 | 90.4 |
| Agent Workflow | 88.2 | 89.7 | 91.2 |
| Code Generation | 79.6 | 84.9 | 90.1 |
| Role Play & Narrative | 94.3 | 94.6 | 95.0 |
| Research & Analysis | 86.3 | 87.3 | 88.3 |
Performance by hardware and tool
Every hardware/tool/quantization combination JonathanColetti-Qwen3.8-27B-Uncensored-GGUF-Q5_K_M 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 | Q5_K_M | 50.8 tok/s | 40.4 tok/s | 31.0 GB | 262,144 tokens | 89.1 | 2 |
Benchmark runs
All 2 JonathanColetti-Qwen3.8-27B-Uncensored-GGUF-Q5_K_M runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is JonathanColetti-Qwen3.8-27B-Uncensored-GGUF-Q5_K_M good for coding?
- In our benchmarks, JonathanColetti-Qwen3.8-27B-Uncensored-GGUF-Q5_K_M scores 84.9/100 for coding. It runs at about 40.4 tok/s, so if you want more speed, local-qwen is faster (~60.2 tok/s) and still scores well for coding (84.8/100).
- Is JonathanColetti-Qwen3.8-27B-Uncensored-GGUF-Q5_K_M good for agentic (tool-using) tasks?
- In our benchmarks, JonathanColetti-Qwen3.8-27B-Uncensored-GGUF-Q5_K_M scores 89.7/100 for agentic workflows. It runs at about 40.4 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 JonathanColetti-Qwen3.8-27B-Uncensored-GGUF-Q5_K_M for local inference?
- Across 2 community benchmark runs, JonathanColetti-Qwen3.8-27B-Uncensored-GGUF-Q5_K_M reaches up to 50.8 tok/s and averages 40.4 tok/s, with the fastest results on NVIDIA RTX A6000.
- How much memory does JonathanColetti-Qwen3.8-27B-Uncensored-GGUF-Q5_K_M need?
- The leanest observed configuration used about 31.0 GB of memory (quantizations tested: Q5_K_M).
- Which tools have been used to run JonathanColetti-Qwen3.8-27B-Uncensored-GGUF-Q5_K_M?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.