Benchmark result
qwen3.8-27b-splash on Apple M5 Max — 74.4 tok/s
Measured with LM Studio on October 10, 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
4th fastest of 11 runs of this model on Apple M5 Max · median 73.0 tok/s.
sort
- 176.6tok/s75.3LM Studio · 4bit
- 276.3tok/s84.7LM Studio · 4bit
- 375.3tok/s84.6LM Studio · 4bit
- 474.4tok/s84.3LM Studio · 4bitthis run
- 573.1tok/s73.8LM Studio · 4bit
- 673.0tok/s85.0LM Studio · 4bit
- 772.7tok/s85.2LM Studio
- 872.4tok/s74.5LM Studio · 4bit
- 971.5tok/s86.0LM Studio · 4bit
- 1071.2tok/s83.4LM Studio · 4bit
- 1171.1tok/s81.6LM Studio
Reproduce this run
toolLM Studio
modelqwen3.8-27b-splash (4bit)
context65,536
clientv0.4.85+101
Set those in LM Studio, then:
llm-benchmark benchmark --model "qwen3.8-27b-splash" --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.