Model benchmarks
qwen3.6-dev:35b local LLM performance
As of November 2026, qwen3.6-dev:35b runs at up to 34.2 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
35.5B
Peak speed
34.2 tok/s
Average speed
33.2 tok/s
Avg PP
279.9 tok/s
Min memory
42.1 GB
Max context
65,536 tokens
Avg output / run
20,049 tokens
Avg runtime / run
10m 37s
Avg quality
74.7
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 | 69.8 | 74.7 | 79.6 |
| Agent Workflow | 62.2 | 70.0 | 77.8 |
| Code Generation | 68.4 | 73.2 | 77.9 |
| Role Play & Narrative | 83.5 | 83.8 | 84.0 |
| Research & Analysis | 65.1 | 71.9 | 78.6 |
Performance by hardware and tool
Every hardware/tool/quantization combination qwen3.6-dev:35b has been benchmarked on, ranked by peak token generation speed. Last updated November 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| NVIDIA RTX 2000 Ada Generation Laptop GPU | Ollama | Q4_K_M | 34.2 tok/s | 33.2 tok/s | 42.1 GB | 65,536 tokens | 74.7 | 2 |
Benchmark runs
All 2 qwen3.6-dev:35b runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is qwen3.6-dev:35b good for coding?
- In our benchmarks, qwen3.6-dev:35b scores 73.2/100 for coding. It runs at about 33.2 tok/s, so if you want more speed, local-qwen is faster (~60.2 tok/s) and still scores well for coding (84.8/100).
- Is qwen3.6-dev:35b good for agentic (tool-using) tasks?
- In our benchmarks, qwen3.6-dev:35b scores 70.0/100 for agentic workflows. It runs at about 33.2 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 qwen3.6-dev:35b for local inference?
- Across 2 community benchmark runs, qwen3.6-dev:35b reaches up to 34.2 tok/s and averages 33.2 tok/s, with the fastest results on NVIDIA RTX 2000 Ada Generation Laptop GPU.
- How much memory does qwen3.6-dev:35b need?
- The leanest observed configuration used about 42.1 GB of memory (quantizations tested: Q4_K_M).
- Which tools have been used to run qwen3.6-dev:35b?
- Benchmarks were submitted using Ollama. Results are community-contributed and updated as new runs arrive.