Benchmark result
MiMo-V2.6-Distill-Qwen-9B-Q6_K on 2× NVIDIA GeForce RTX 5060 — 64.2 tok/s
Measured with llama.cpp b11093-fb34fc262 on September 22, 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 2 runs of this model on 2× NVIDIA GeForce RTX 5060 · Multi-GPU · median 58.7 tok/s.
sort
- 164.2tok/s67.0llama.cpp · Q6_Kthis run
- 253.2tok/s73.5llama.cpp · Q6_K
Reproduce this run
toolllama.cpp b11093-fb34fc262
modelMiMo-V2.6-Distill-Qwen-9B-Q6_K (Q6_K)
context262,144
thinkingon
samplingtemperature 0.699999988079071 · top_p 0.8999999761581421 · top_k 20 · min_p 0.05000000074505806 · presence_penalty 0 · repeat_penalty 1.0499999523162842
clientv0.4.80+98
Set those in llama.cpp, then:
llm-benchmark benchmark --model "MiMo-V2.6-Distill-Qwen-9B-Q6_K" --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.