Benchmark result

qwen3.8-27b-splash on Apple M4 Max — 73.0 tok/s

Measured with LM Studio on October 3, 2026.

ShareRedditX

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

56th fastest of 97 runs of this model on Apple M4 Max · median 73.6 tok/s.

sort
Compare the nearest 5 →

qwen3.8-27b-splash on other hardware

Reproduce this run

toolLM Studio
modelqwen3.8-27b-splash (4bit)
context65,536
clientv0.4.87+102

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.