Benchmark result
LFM2.5-24B-A2B-Q4_K_M on AMD Radeon 680M — 19.7 tok/s
Measured with llama.cpp b10787-73bfe6cc4 on October 30, 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 b10787-73bfe6cc4
modelLFM2.5-24B-A2B-Q4_K_M (Q4_K_M)
context8,192
thinkingoff
samplingtemperature 0.10000000149011612 · top_p 0.949999988079071 · top_k 50 · min_p 0.05000000074505806 · presence_penalty 0 · repeat_penalty 1.0499999523162842
clientv0.4.98+102
Set those in llama.cpp, then:
llm-benchmark benchmark --model "LFM2.5-24B-A2B-Q4_K_M" --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.