Model benchmarks
Qwen3.8:27B/MTP@180K local LLM performance
As of August 2026, Qwen3.8:27B/MTP@180K runs at up to 53.0 tok/s for local inference (best of 2 community benchmark runs across 0 GPUs).
LM Studio
Model size
27B
Peak speed
53.0 tok/s
Average speed
47.6 tok/s
Min memory
13.2 GB
Max context
14,274 tokens
Best quality
93.3
Benchmark runs
2
GPUs tested
0
Performance by hardware and tool
Every hardware/tool/quantization combination Qwen3.8:27B/MTP@180K has been benchmarked on, ranked by peak token generation speed. Last updated August 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| CPU only | LM Studio | — | 53.0 tok/s | 47.6 tok/s | 13.2 GB | 14,274 tokens | 57.4 | 2 |
Frequently asked questions
- How fast is Qwen3.8:27B/MTP@180K for local inference?
- Across 2 community benchmark runs, Qwen3.8:27B/MTP@180K reaches up to 53.0 tok/s and averages 47.6 tok/s.
- How much memory does Qwen3.8:27B/MTP@180K need?
- The leanest observed configuration used about 13.2 GB of memory.
- Which tools have been used to run Qwen3.8:27B/MTP@180K?
- Benchmarks were submitted using LM Studio. Results are community-contributed and updated as new runs arrive.