Model benchmarks
Swift-1.5-Qwen3.8-27b-oQ8e-mtp local LLM performance
As of September 2026, Swift-1.5-Qwen3.8-27b-oQ8e-mtp runs at up to 34.8 tok/s for local inference (best of 5 community benchmark runs across 1 GPU).
Model size
27B
Peak speed
34.8 tok/s
Average speed
31.3 tok/s
Avg PP
409.4 tok/s
Min memory
28.6 GB
Max context
262,144 tokens
Avg output / run
49,461 tokens
Avg runtime / run
27m 9s
Avg quality
85.2
Benchmark runs
5
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 5 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 | 83.2 | 85.2 | 87.1 |
| Agent Workflow | 79.3 | 85.8 | 90.2 |
| Code Generation | 72.1 | 79.9 | 85.3 |
| Role Play & Narrative | 82.6 | 89.5 | 93.8 |
| Research & Analysis | 83.0 | 85.6 | 88.3 |
Performance by hardware and tool
Every hardware/tool/quantization combination Swift-1.5-Qwen3.8-27b-oQ8e-mtp 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 M5 Max | oMLX | oQ8e | 34.8 tok/s | 31.3 tok/s | 28.6 GB | 262,144 tokens | 85.2 | 5 |
Benchmark runs
All 5 Swift-1.5-Qwen3.8-27b-oQ8e-mtp runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is Swift-1.5-Qwen3.8-27b-oQ8e-mtp good for coding?
- In our benchmarks, Swift-1.5-Qwen3.8-27b-oQ8e-mtp scores 79.9/100 for coding. It runs at about 31.3 tok/s, so if you want more speed, Huihui-Qwen3.8-27B-abliterated-Swift-1.5-GSQ-RCO-IQ3_S-mtp is faster (~43.5 tok/s) and still scores well for coding (86.6/100).
- Is Swift-1.5-Qwen3.8-27b-oQ8e-mtp good for agentic (tool-using) tasks?
- In our benchmarks, Swift-1.5-Qwen3.8-27b-oQ8e-mtp scores 85.8/100 for agentic workflows. It runs at about 31.3 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 Swift-1.5-Qwen3.8-27b-oQ8e-mtp for local inference?
- Across 5 community benchmark runs, Swift-1.5-Qwen3.8-27b-oQ8e-mtp reaches up to 34.8 tok/s and averages 31.3 tok/s, with the fastest results on Apple M5 Max.
- How much memory does Swift-1.5-Qwen3.8-27b-oQ8e-mtp need?
- The leanest observed configuration used about 28.6 GB of memory (quantizations tested: oQ8e).
- Which tools have been used to run Swift-1.5-Qwen3.8-27b-oQ8e-mtp?
- Benchmarks were submitted using oMLX. Results are community-contributed and updated as new runs arrive.