Benchmark result
unsloth/Devstral-Small-2-24B-Instruct-2512-GGUF:UD-Q3_K_XL on AMD Radeon RX 9070/9070 XT/9070 GRE — 36.2 tok/s
Measured with llama.cpp b11388-050439614 on October 5, 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 b11388-050439614
modelunsloth/Devstral-Small-2-24B-Instruct-2512-GGUF:UD-Q3_K_XL (Q3_K)
context32,768
think limitnone
samplingtemperature 0.15000000596046448 · top_p 0.949999988079071 · top_k 40 · min_p 0.05000000074505806 · presence_penalty 0 · repeat_penalty 1
clientv0.4.87+102
Set those in llama.cpp, then:
llm-benchmark benchmark --model "unsloth/Devstral-Small-2-24B-Instruct-2512-GGUF:UD-Q3_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.