Benchmark result
Qwen3.8-9B-Distill-uncensored-heretic.i1-IQ3_M on Tesla P100-PCIE-16GB — 34.2 tok/s
Measured with llama.cpp b1-5266f24 on September 21, 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
2nd fastest of 3 runs of this model on Tesla P100-PCIE-16GB · median 34.2 tok/s.
sort
- 134.3tok/s51.8llama.cpp
- 234.2tok/s62.6llama.cppthis run
- 334.2tok/s64.9llama.cpp
Reproduce this run
toolllama.cpp b1-5266f24
modelQwen3.8-9B-Distill-uncensored-heretic.i1-IQ3_M
context65,536
thinkingoff
samplingtemperature 0.800000011920929 · top_p 0.949999988079071 · top_k 40 · min_p 0.05000000074505806 · presence_penalty 0 · repeat_penalty 1
clientv0.4.71+98
Set those in llama.cpp, then:
llm-benchmark benchmark --model "Qwen3.8-9B-Distill-uncensored-heretic.i1-IQ3_M" --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.