Benchmark result
unsloth/Qwen3.8-27B-GGUF on Apple M5 Max — 29.5 tok/s
Measured with llama.cpp b11429-d81235049 on October 15, 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
3rd fastest of 4 runs of this model on Apple M5 Max · median 34.6 tok/s.
sort
- 140.4tok/s84.2Unsloth Studio · UD-Q6_K_XL · KV f16
- 239.8tok/s83.2Unsloth Studio · UD-Q6_K_XL · KV q8_0
- 329.5tok/s86.1llama.cppthis run
- 429.4tok/s87.4llama.cpp
unsloth/Qwen3.8-27B-GGUF on other hardware
Reproduce this run
toolllama.cpp b11429-d81235049
modelunsloth/Qwen3.8-27B-GGUF
context65,536
thinkingon
effortmedium
samplingtemperature 1 · top_p 0.949999988079071 · top_k 20 · min_p 0.05000000074505806 · presence_penalty 0 · repeat_penalty 1
clientv0.4.85+101
Set those in llama.cpp, then:
llm-benchmark benchmark --model "unsloth/Qwen3.8-27B-GGUF" --tool "llama.cpp"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.