Model benchmarks
nvidia/Qwen3.8-27B-NVFP4 local LLM performance
As of October 2026, nvidia/Qwen3.8-27B-NVFP4 runs at up to 75.6 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
27B
Peak speed
75.6 tok/s
Average speed
75.5 tok/s
Avg PP
82.8 tok/s
Min memory
16.6 GB
Max context
131,072 tokens
Avg output / run
12,144 tokens
Avg runtime / run
2m 46s
Avg quality
75.1
Benchmark runs
2
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 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 | 73.7 | 75.1 | 76.5 |
| Agent Workflow | 65.1 | 67.5 | 69.8 |
| Code Generation | 82.2 | 83.5 | 84.9 |
| Role Play & Narrative | 66.6 | 71.4 | 76.1 |
| Research & Analysis | 75.4 | 78.2 | 81.0 |
Performance by hardware and tool
Every hardware/tool/quantization combination nvidia/Qwen3.8-27B-NVFP4 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 | OpenAI-compatible | — | 75.6 tok/s | 75.5 tok/s | 16.6 GB | 131,072 tokens | 75.1 | 2 |
Benchmark runs
All 2 nvidia/Qwen3.8-27B-NVFP4 runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is nvidia/Qwen3.8-27B-NVFP4 good for coding?
- In our benchmarks, nvidia/Qwen3.8-27B-NVFP4 scores 83.5/100 for coding, among the top 3 for coding on consumer hardware (≤24 GB VRAM). It runs at about 75.5 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 nvidia/Qwen3.8-27B-NVFP4 good for agentic (tool-using) tasks?
- In our benchmarks, nvidia/Qwen3.8-27B-NVFP4 scores 67.5/100 for agentic workflows. It runs at about 75.5 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 nvidia/Qwen3.8-27B-NVFP4 for local inference?
- Across 2 community benchmark runs, nvidia/Qwen3.8-27B-NVFP4 reaches up to 75.6 tok/s and averages 75.5 tok/s, with the fastest results on Apple M5 Max.
- How much memory does nvidia/Qwen3.8-27B-NVFP4 need?
- The leanest observed configuration used about 16.6 GB of memory.
- Which tools have been used to run nvidia/Qwen3.8-27B-NVFP4?
- Benchmarks were submitted using OpenAI-compatible. Results are community-contributed and updated as new runs arrive.