Model benchmarks
Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUF-RVN-Q5_K_M-mtp local LLM performance
As of October 2026, Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUF-RVN-Q5_K_M-mtp runs at up to 29.7 tok/s for local inference (best of 3 community benchmark runs across 1 GPU).
Model size
27B
Peak speed
29.7 tok/s
Average speed
29.7 tok/s
Avg PP
562.8 tok/s
Min memory
30.4 GB
Max context
262,144 tokens
Avg output / run
31,489 tokens
Avg runtime / run
18m
Avg quality
84.2
Benchmark runs
3
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 3 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 | 83.0 | 84.2 | 85.6 |
| Agent Workflow | 81.5 | 83.9 | 85.7 |
| Code Generation | 73.1 | 78.5 | 82.4 |
| Role Play & Narrative | 84.8 | 88.4 | 91.6 |
| Research & Analysis | 82.3 | 86.0 | 89.3 |
Performance by hardware and tool
Every hardware/tool/quantization combination Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUF-RVN-Q5_K_M-mtp 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 | 29.7 tok/s | 29.7 tok/s | 30.4 GB | 262,144 tokens | 84.2 | 3 |
Benchmark runs
All 3 Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUF-RVN-Q5_K_M-mtp runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUF-RVN-Q5_K_M-mtp good for coding?
- In our benchmarks, Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUF-RVN-Q5_K_M-mtp scores 78.5/100 for coding. It runs at about 29.7 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 Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUF-RVN-Q5_K_M-mtp good for agentic (tool-using) tasks?
- In our benchmarks, Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUF-RVN-Q5_K_M-mtp scores 83.9/100 for agentic workflows. It runs at about 29.7 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-Heretic-Abliterated-Uncensored-GGUF-RVN-Q5_K_M-mtp for local inference?
- Across 3 community benchmark runs, Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUF-RVN-Q5_K_M-mtp reaches up to 29.7 tok/s and averages 29.7 tok/s, with the fastest results on NVIDIA RTX A6000.
- How much memory does Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUF-RVN-Q5_K_M-mtp need?
- The leanest observed configuration used about 30.4 GB of memory (quantizations tested: Q5_K_M).
- Which tools have been used to run Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUF-RVN-Q5_K_M-mtp?
- Benchmarks were submitted using llama.cpp. Results are community-contributed and updated as new runs arrive.