Model benchmarks
qwen3.5-4b@q6_k local LLM performance
As of October 2026, qwen3.5-4b@q6_k runs at up to 29.6 tok/s for local inference (best of 3 community benchmark runs across 1 GPU).
Model size
4B
Peak speed
29.6 tok/s
Average speed
28.9 tok/s
Avg PP
874.4 tok/s
Min memory
6.0 GB
Max context
32,768 tokens
Avg output / run
17,005 tokens
Avg runtime / run
10m 19s
Avg quality
71.5
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 | 68.1 | 71.5 | 74.9 |
| Agent Workflow | 72.6 | 73.9 | 75.5 |
| Code Generation | 49.3 | 54.8 | 62.7 |
| Role Play & Narrative | 72.6 | 79.1 | 83.8 |
| Research & Analysis | 76.7 | 78.1 | 80.0 |
Performance by hardware and tool
Every hardware/tool/quantization combination qwen3.5-4b@q6_k 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 |
|---|---|---|---|---|---|---|---|---|
| NVIDIA GeForce RTX 3060 Laptop GPU | Bionic | Q6_K | 29.6 tok/s | 28.9 tok/s | 6.0 GB | 32,768 tokens | 71.5 | 3 |
Benchmark runs
All 3 qwen3.5-4b@q6_k runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is qwen3.5-4b@q6_k good for coding?
- In our benchmarks, qwen3.5-4b@q6_k scores 54.8/100 for coding. It runs at about 28.9 tok/s, so if you want more speed, Qwen3.8-27B-Turbo-Brilliance-50X-RI-GSQ-RCO-IQ3_XXS-mtp is faster (~37.5 tok/s) and still scores well for coding (83.8/100).
- Is qwen3.5-4b@q6_k good for agentic (tool-using) tasks?
- In our benchmarks, qwen3.5-4b@q6_k scores 73.9/100 for agentic workflows. It runs at about 28.9 tok/s, so if you want more speed, muse-glimmer:latest is faster (~33.3 tok/s) and still scores well for agentic workflows (91.6/100).
- How fast is qwen3.5-4b@q6_k for local inference?
- Across 3 community benchmark runs, qwen3.5-4b@q6_k reaches up to 29.6 tok/s and averages 28.9 tok/s, with the fastest results on NVIDIA GeForce RTX 3060 Laptop GPU.
- How much memory does qwen3.5-4b@q6_k need?
- The leanest observed configuration used about 6.0 GB of memory (quantizations tested: Q6_K).
- Which tools have been used to run qwen3.5-4b@q6_k?
- Benchmarks were submitted using Bionic. Results are community-contributed and updated as new runs arrive.