Model benchmarks
qwen3.5:latest local LLM performance
As of July 2026, qwen3.5:latest runs at up to 69.5 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).
Model size
9.7B
Peak speed
69.5 tok/s
Average speed
52.7 tok/s
Min memory
5.5 GB
Max context
8.192 tokens
Avg quality
32.7
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 | 26.5 | 32.7 | 38.9 |
| Agent Workflow | 4.5 | 45.2 | 85.9 |
| Code Generation | 1.1 | 2.3 | 3.4 |
| Role Play & Narrative | 55.6 | 59.8 | 64.0 |
| Research & Analysis | 2.3 | 23.4 | 44.6 |
Performance by hardware and tool
Every hardware/tool/quantization combination qwen3.5:latest has been benchmarked on, ranked by peak token generation speed. Last updated July 2026.
| Hardware | Tool | Quant | Best tok/s | Avg tok/s | Memory | Context | Quality | Runs |
|---|---|---|---|---|---|---|---|---|
| CPU only | Ollama | Q4_K_M | 69.5 tok/s | 69.5 tok/s | 5.5 GB | 8.192 tokens | 39.6 | 1 |
| Apple M4 Pro | Ollama | Q4_K_M | 35.9 tok/s | 35.9 tok/s | 5.5 GB | 8.192 tokens | 25.8 | 1 |
Benchmark runs
All 2 qwen3.5:latest runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.
Frequently asked questions
- Is qwen3.5:latest good for coding?
- In our benchmarks, qwen3.5:latest scores 2.3/100 for coding. It runs at about 52.7 tok/s, so if you want more speed, unsloth/Qwen3.8-27B-GGUF:IQ3_S is faster (~104.3 tok/s) and still scores well for coding (83.0/100).
- Is qwen3.5:latest good for agentic (tool-using) tasks?
- In our benchmarks, qwen3.5:latest scores 45.2/100 for agentic workflows. It runs at about 52.7 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 qwen3.5:latest for local inference?
- Across 2 community benchmark runs, qwen3.5:latest reaches up to 69.5 tok/s and averages 52.7 tok/s.
- How much memory does qwen3.5:latest need?
- The leanest observed configuration used about 5.5 GB of memory (quantizations tested: Q4_K_M).
- Which tools have been used to run qwen3.5:latest?
- Benchmarks were submitted using Ollama. Results are community-contributed and updated as new runs arrive.