Benchmark result

ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF on NVIDIA GeForce RTX 4070 — 41.6 tok/s

Measured with llama.cpp manual on November 1, 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

2nd fastest of 6 runs of this model on NVIDIA GeForce RTX 4070 · median 39.7 tok/s.

sort

141.8tok/s74.2Unsloth Studio · IQ2_XS · KV q4_0
241.6tok/s84.7llama.cppthis run
339.8tok/s76.9llama.cpp
439.6tok/s82.0llama.cpp
538.8tok/s81.9llama.cpp
638.5tok/s87.7llama.cpp

Compare the nearest 5 →

ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF on other hardware

123.0tok/s78.2NVIDIA GeForce RTX 5090

Reproduce this run

toolllama.cpp manual

modelISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF

context32,768

thinkingon

effortmedium

think limit16,384 tokens

samplingtemperature 1 · top_p 0.949999988079071 · top_k 20 · min_p 0 · presence_penalty 0 · repeat_penalty 1

clientv0.4.98+102

Set those in llama.cpp, then:

llm-benchmark benchmark --model "ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF" --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.