Benchmark result
qwen3.5-4b-mlx on Apple M1 Max — 78.5 tok/s
Measured with LM Studio on November 3, 2026.
Support llm-bench.io
llm-bench.io is free. If it helped you, I'd appreciate your support to cover the infrastructure costs.
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
2nd fastest of 2 runs of this model on Apple M1 Max · median 81.4 tok/s.
sort
- 184.2tok/s54.4LM Studio · 4bit
- 278.5tok/s60.7LM Studio · 4bitthis run
qwen3.5-4b-mlx on other hardware
Reproduce this run
toolLM Studio
modelqwen3.5-4b-mlx (4bit)
context65,536
clientv0.4.98+102
Set those in LM Studio, then:
llm-benchmark benchmark --model "qwen3.5-4b-mlx" --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.