Benchmark result
qwen3.8-27b on 2× NVIDIA A10G — 79.4 tok/s
Measured with ninfer manual on November 3, 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
Reproduce this run
toolninfer manual
modelqwen3.8-27b
context65,536
thinkingon
clientv0.4.98+102
Set those in ninfer, then:
llm-benchmark benchmark --model "qwen3.8-27b" --tool "ninfer"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.