Benchmark result
qwen3.8-27b-gsq-rco@iq2_xs on AMD Radeon RX 7900 GRE — 33.9 tok/s
Measured with LM Studio on October 8, 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
1st fastest of 3 runs of this model on AMD Radeon RX 7900 GRE · median 31.9 tok/s.
sort
- 133.9tok/s73.4LM Studio · IQ2_XSthis run
- 231.9tok/s66.2LM Studio · IQ2_XS
- 328.0tok/s87.3LM Studio · IQ2_XS
Reproduce this run
toolLM Studio
modelqwen3.8-27b-gsq-rco@iq2_xs (IQ2_XS)
context32,768
thinkingoff
clientv0.4.89+102
Set those in LM Studio, then:
llm-benchmark benchmark --model "qwen3.8-27b-gsq-rco@iq2_xs" --tool "LM Studio"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.