Benchmark result
Kwaipilot-KAT-Coder on Tesla V100-PCIE-32GB +2 more — 81.2 tok/s
Measured with llama.cpp manual on November 9, 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
2nd fastest of 5 runs of this model on Tesla V100-PCIE-32GB +2 more · Multi-GPU · median 81.2 tok/s.
sort
- 181.4tok/s66.7llama.cpp
- 281.2tok/s83.0llama.cppthis run
- 381.2tok/s82.1OpenAI-compatible
- 480.9tok/s82.6OpenAI-compatible
- 580.4tok/s60.0llama.cpp
Kwaipilot-KAT-Coder on other hardware
Reproduce this run
toolllama.cpp manual
modelKwaipilot-KAT-Coder
context32,768
think limit8,096 tokens
samplingtemperature 1 · top_p 0.949999988079071 · top_k 20 · min_p 0 · presence_penalty 1.5 · repeat_penalty 1
clientv0.4.99+102
Set those in llama.cpp, then:
llm-benchmark benchmark --model "Kwaipilot-KAT-Coder" --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.