Benchmark result
Qwen3.8-27B-EfficientThink-VAL-APEX-I-NanoPlus-V3 on 2× NVIDIA GeForce RTX 5060 — 38.2 tok/s
Measured with llama.cpp b11609-fcac19079 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.
Reproduce this run
toolllama.cpp b11609-fcac19079
modelQwen3.8-27B-EfficientThink-VAL-APEX-I-NanoPlus-V3
context131,072
thinkingon
effortmedium
think limitnone
samplingtemperature 0.6000000238418579 · top_p 0.8500000238418579 · top_k 20 · min_p 0.10000000149011612 · presence_penalty 0 · repeat_penalty 1.0499999523162842
clientv0.4.93+102
Set those in llama.cpp, then:
llm-benchmark benchmark --model "Qwen3.8-27B-EfficientThink-VAL-APEX-I-NanoPlus-V3" --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.