Benchmark result
gpt-oss-20b-UD-Q8_K_XL on 2× NVIDIA GeForce RTX 5060 — 176.2 tok/s
Measured with llama.cpp b10884-434ddbbc0 on September 9, 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 5 runs of this model on 2× NVIDIA GeForce RTX 5060 · Multi-GPU · median 114.5 tok/s.
sort
- 1176.2tok/s77.2llama.cpp · Q8_Kthis run
- 2159.6tok/s73.7llama.cpp · Q8_K
- 3114.5tok/s68.2llama.cpp · Q8_K
- 4109.4tok/s75.9llama.cpp · Q8_K
- 587.5tok/s71.8llama.cpp · Q8_K
Reproduce this run
toolllama.cpp b10884-434ddbbc0
modelgpt-oss-20b-UD-Q8_K_XL (Q8_K)
context131,072
thinkingon
effortmedium (model default)
samplingtemperature 0.20000000298023224 · 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 "gpt-oss-20b-UD-Q8_K_XL" --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.