Model benchmarks
mtplx-qwen38-27b-optimized-speed local LLM performance
As of September 2026, mtplx-qwen38-27b-optimized-speed runs at up to 45.5 tok/s for local inference (best of 2 community benchmark runs across 2 GPUs).
Model size
27B
Peak speed
45.5 tok/s
Average speed
43.2 tok/s
Min memory
n/a
Max context
128,000 tokens
Avg output / run
82,669 tokens
Avg runtime / run
32m 36s
Avg quality
86.1
Benchmark runs
2
GPUs tested
2
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 | 85.8 | 86.1 | 86.4 |
| Agent Workflow | 87.5 | 89.3 | 91.1 |
| Code Generation | 70.9 | 73.3 | 75.7 |
| Role Play & Narrative | 93.8 | 94.0 | 94.2 |
| Research & Analysis | 87.0 | 87.7 | 88.4 |
Performance by hardware and tool
Every hardware/tool/quantization combination mtplx-qwen38-27b-optimized-speed has been benchmarked on, ranked by peak token generation speed. Last updated September 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| Apple M4 Max | vLLM | — | 45.5 tok/s | 45.5 tok/s | n/a | 128,000 tokens | 86.4 | 1 |
| Apple M5 Max | vLLM | — | 40.8 tok/s | 40.8 tok/s | n/a | 128,000 tokens | 85.8 | 1 |
Frequently asked questions
- Is mtplx-qwen38-27b-optimized-speed good for coding?
- In our benchmarks, mtplx-qwen38-27b-optimized-speed scores 73.3/100 for coding. It runs at about 43.2 tok/s, so if you want more speed, Gemma4:E2B/QAT-MTP@131K is faster (~303.9 tok/s) and still scores well for coding (66.9/100).
- Is mtplx-qwen38-27b-optimized-speed good for agentic (tool-using) tasks?
- In our benchmarks, mtplx-qwen38-27b-optimized-speed scores 89.3/100 for agentic workflows. It runs at about 43.2 tok/s, so if you want more speed, mudler/Ornith-1.5-35B-A3B-APEX-MTP-GGUF is faster (~190.9 tok/s) and still scores well for agentic workflows (85.7/100).
- How fast is mtplx-qwen38-27b-optimized-speed for local inference?
- Across 2 community benchmark runs, mtplx-qwen38-27b-optimized-speed reaches up to 45.5 tok/s and averages 43.2 tok/s, with the fastest results on Apple M4 Max.
- How much memory does mtplx-qwen38-27b-optimized-speed need?
- The leanest observed configuration used about n/a of memory.
- Which tools have been used to run mtplx-qwen38-27b-optimized-speed?
- Benchmarks were submitted using vLLM. Results are community-contributed and updated as new runs arrive.