Benchmark result
Kwaipilot-KAT-Coder_t1p95k20m0 on Tesla V100-PCIE-32GB +2 more — 99.3 tok/s
Measured with llama.cpp b12231-a0f8e8ac2 on November 10, 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
3rd fastest of 3 runs of this model on Tesla V100-PCIE-32GB +2 more · Multi-GPU · median 101.1 tok/s.
sort
- 1101.2tok/s78.3llama.cpp
- 2101.1tok/s76.4llama.cpp
- 399.3tok/s82.8llama.cppthis run
Reproduce this run
toolllama.cpp b12231-a0f8e8ac2
modelKwaipilot-KAT-Coder_t1p95k20m0
context32,768
think limit8,096 tokens
samplingtemperature 1 · top_p 0.949999988079071 · top_k 20 · min_p 0 · presence_penalty 0 · repeat_penalty 1
clientv0.4.99+102
Set those in llama.cpp, then:
llm-benchmark benchmark --model "Kwaipilot-KAT-Coder_t1p95k20m0" --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.