Benchmark result
Qwen3.6-35B-A3B-UD-Q5_K_XL on Radeon Instinct MI60 — 58.4 tok/s
Measured with llama.cpp b10384-6b517e82a on September 4, 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.
Reproduce this run
toolllama.cpp b10384-6b517e82a
modelQwen3.6-35B-A3B-UD-Q5_K_XL (Q5_K)
context8,192
thinkingoff
samplingtemperature 0.5 · top_p 0.800000011920929 · top_k 20 · min_p 0 · presence_penalty 1.5 · repeat_penalty 1
clientv0.4.54+97
Set those in llama.cpp, then:
llm-benchmark benchmark --model "Qwen3.6-35B-A3B-UD-Q5_K_XL" --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.