Benchmark result
Qwen3.5:4B/MTP@262K — 199.2 tok/s
Measured with LM Studio on August 21, 2026.
What the model built
Open full screen →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
Qwen3.5:4B/MTP@262K on other hardware
- 362.6tok/s22.4Unknown hardware
- 199.2tok/s41.0Unknown hardwarethis run
Reproduce this run
toolLM Studio
modelQwen3.5:4B/MTP@262K
context16,384
clientv0.4.40+97
Set those in LM Studio, then:
llm-benchmark benchmark --model "Qwen3.5:4B/MTP@262K" --tool "LM Studio"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.