Benchmark result

qwen3.5:9b on NVIDIA RTX 2000 Ada Generation Laptop GPU — 14.1 tok/s

Measured with Ollama 0.40.13 on October 31, 2026.

What the model built

The coding scenario asks for a playable game in a single HTML file. This is exactly what the model returned, unedited; the frame only adds a line that tells this page how tall it is.

How this run compares

2nd fastest of 2 runs of this model on NVIDIA RTX 2000 Ada Generation Laptop GPU · median 20.9 tok/s.

sort

127.7tok/s61.6Ollama · Q4_K_M
214.1tok/s72.3Ollama · Q4_K_Mthis run

Compare the nearest 5 →

qwen3.5:9b on other hardware

Reproduce this run

toolOllama 0.40.13

modelqwen3.5:9b (Q4_K_M)

context65,536

samplingtemperature 1 · top_p 0.95 · top_k 20 · presence_penalty 1.5

clientv0.4.98+102

Set those in Ollama, then:

llm-benchmark benchmark --model "qwen3.5:9b" --tool "Ollama"

The client prompts for context length and thinking mode, and for the KV cache dtype on Unsloth Studio. Sampling is left at the model default — the values above are what the tool reported using, not overrides the client sent.