Model benchmarks
Qwopus3.8-27B-Flash-V2-oQ4e-mtp local LLM performance
As of October 2026, Qwopus3.8-27B-Flash-V2-oQ4e-mtp runs at up to 68.1 tok/s for local inference (best of 3 community benchmark runs across 1 GPU).
Model size
27B
Peak speed
68.1 tok/s
Average speed
66.9 tok/s
Avg PP
1231.8 tok/s
Min memory
16.9 GB
Max context
65,536 tokens
Avg output / run
18,577 tokens
Avg runtime / run
4m 37s
Avg quality
74.3
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 | 69.8 | 74.3 | 77.4 |
| Agent Workflow | 71.3 | 76.8 | 80.9 |
| Code Generation | 52.6 | 56.9 | 61.1 |
| Role Play & Narrative | 74.7 | 84.3 | 92.1 |
| Research & Analysis | 71.4 | 79.3 | 86.7 |
Performance by hardware and tool
Every hardware/tool/quantization combination Qwopus3.8-27B-Flash-V2-oQ4e-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 |
|---|---|---|---|---|---|---|---|---|
| Apple M5 Max | oMLX | oQ4e | 68.1 tok/s | 66.9 tok/s | 16.9 GB | 65,536 tokens | 74.3 | 3 |
Benchmark runs
All 3 Qwopus3.8-27B-Flash-V2-oQ4e-mtp runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is Qwopus3.8-27B-Flash-V2-oQ4e-mtp good for coding?
- In our benchmarks, Qwopus3.8-27B-Flash-V2-oQ4e-mtp scores 56.9/100 for coding. It runs at about 66.9 tok/s, so if you want more speed, Nex-N2.5-mini-IQ3_XXS is faster (~91.4 tok/s) and still scores well for coding (81.6/100).
- Is Qwopus3.8-27B-Flash-V2-oQ4e-mtp good for agentic (tool-using) tasks?
- In our benchmarks, Qwopus3.8-27B-Flash-V2-oQ4e-mtp scores 76.8/100 for agentic workflows. It runs at about 66.9 tok/s, so if you want more speed, Nex-N2.5-mini-APEX-Mini is faster (~88.2 tok/s) and still scores well for agentic workflows (91.3/100).
- How fast is Qwopus3.8-27B-Flash-V2-oQ4e-mtp for local inference?
- Across 3 community benchmark runs, Qwopus3.8-27B-Flash-V2-oQ4e-mtp reaches up to 68.1 tok/s and averages 66.9 tok/s, with the fastest results on Apple M5 Max.
- How much memory does Qwopus3.8-27B-Flash-V2-oQ4e-mtp need?
- The leanest observed configuration used about 16.9 GB of memory (quantizations tested: oQ4e).
- Which tools have been used to run Qwopus3.8-27B-Flash-V2-oQ4e-mtp?
- Benchmarks were submitted using oMLX. Results are community-contributed and updated as new runs arrive.