Benchmark result

Qwen3.8-27B-GSQ-RCO-IQ3_S-mtp on 2× NVIDIA GeForce RTX 5060 — 27.1 tok/s

Measured with llama.cpp b10918-82d6bb284 on September 12, 2026.

What the model built

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

Qwen3.8-27B-GSQ-RCO-IQ3_S-mtp on other hardware

35.0tok/s87.0AMD Radeon RX 9070/9070 XT/9070 GRE

Reproduce this run

toolllama.cpp b10918-82d6bb284

modelQwen3.8-27B-GSQ-RCO-IQ3_S-mtp

context98,304

thinkingon

effortxhigh (model default)

samplingtemperature 0.6000000238418579 · top_p 0.8999999761581421 · top_k 0 · min_p 0.05000000074505806 · presence_penalty 0 · repeat_penalty 1

clientv0.4.59+97

Set those in llama.cpp, then:

llm-benchmark benchmark --model "Qwen3.8-27B-GSQ-RCO-IQ3_S-mtp" --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.