Benchmark result
bonsai-2-27b on NVIDIA GeForce RTX 2070 SUPER — 39.2 tok/s
Measured with llama-server manual on October 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 NVIDIA GeForce RTX 2070 SUPER · median 38.7 tok/s.
sort
- 139.2tok/s—llama-serverthis run
- 238.3tok/s—llama-server (Bonsai 2 27B PTQ1_0)
Reproduce this run
toolllama-server manual
modelbonsai-2-27b
context8,192
clientv0.4.96+102
Set those in llama-server, then:
llm-benchmark benchmark --model "bonsai-2-27b" --tool "llama-server"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.