Benchmark result
Qwen3.8-27B-IQ3_XXS on AMD Radeon RX 7900 XTX — 58.0 tok/s
Measured with llama.cpp b182-c1d0e7a on August 28, 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 2 runs of this model on AMD Radeon RX 7900 XTX · median 59.6 tok/s.
sort
- 161.1tok/s85.6llama.cpp
- 258.0tok/s83.1llama.cppthis run
Qwen3.8-27B-IQ3_XXS on other hardware
Reproduce this run
toolllama.cpp b182-c1d0e7a
modelQwen3.8-27B-IQ3_XXS
context65,536
thinkingon
clientv0.4.44+97
Set those in llama.cpp, then:
llm-benchmark benchmark --model "Qwen3.8-27B-IQ3_XXS" --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.