Benchmark result

peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q6_K_XL on NVIDIA GeForce RTX 4070 — 157.1 tok/s

Measured with llama.cpp b11643-3bf0e0594 on October 15, 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

peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q6_K_XL on other hardware

61.3tok/s56.4NVIDIA GeForce RTX 2060 · Q6_K

Reproduce this run

toolllama.cpp b11643-3bf0e0594

modelpeculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q6_K_XL (Q6_K)

context131,072

thinkingon

think limitnone

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

clientv0.4.96+102

Set those in llama.cpp, then:

llm-benchmark benchmark --model "peculiar-ragdoll/Sharp-MiniCPM5-2B-GGUF:Q6_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.