Model benchmarks
mtplx-qwen38-27b-optimized-quality local LLM performance
As of August 2026, mtplx-qwen38-27b-optimized-quality runs at up to 36.2 tok/s for local inference (best of 3 community benchmark runs across 2 GPUs).
Model size
27B
Peak speed
36.2 tok/s
Average speed
30.7 tok/s
Min memory
n/a
Max context
128,000 tokens
Avg output / run
37,035 tokens
Avg runtime / run
15m 31s
Avg quality
84.6
Benchmark runs
3
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 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 | 81.0 | 84.6 | 87.9 |
| Agent Workflow | 74.5 | 81.2 | 87.4 |
| Code Generation | 74.7 | 79.5 | 84.8 |
| Role Play & Narrative | 91.1 | 91.8 | 92.7 |
| Research & Analysis | 83.3 | 85.9 | 88.2 |
Performance by hardware and tool
Every hardware/tool/quantization combination mtplx-qwen38-27b-optimized-quality 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 |
|---|---|---|---|---|---|---|---|---|
| Apple M4 Max | vLLM | — | 36.2 tok/s | 35.5 tok/s | n/a | 128,000 tokens | 86.6 | 2 |
| Apple M5 Pro | vLLM | — | 21.0 tok/s | 21.0 tok/s | n/a | 8,192 tokens | 80.5 | 1 |
Frequently asked questions
- Is mtplx-qwen38-27b-optimized-quality good for coding?
- In our benchmarks, mtplx-qwen38-27b-optimized-quality scores 79.5/100 for coding. It runs at about 30.7 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 coding (73.4/100).
- Is mtplx-qwen38-27b-optimized-quality good for agentic (tool-using) tasks?
- In our benchmarks, mtplx-qwen38-27b-optimized-quality scores 81.2/100 for agentic workflows. It runs at about 30.7 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-quality for local inference?
- Across 3 community benchmark runs, mtplx-qwen38-27b-optimized-quality reaches up to 36.2 tok/s and averages 30.7 tok/s, with the fastest results on Apple M4 Max.
- How much memory does mtplx-qwen38-27b-optimized-quality need?
- The leanest observed configuration used about n/a of memory.
- Which tools have been used to run mtplx-qwen38-27b-optimized-quality?
- Benchmarks were submitted using vLLM. Results are community-contributed and updated as new runs arrive.