Benchmark result
qwen36-agent on AMD Radeon Graphics — 97.5 tok/s
Measured with llama.cpp b12192-9f0e5098a on November 20, 2026.
Support llm-bench.io
llm-bench.io is free. If it helped you, I'd appreciate your support to cover the infrastructure costs.
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
qwen36-agent on other hardware
- 97.5tok/s75.9Unknown hardwarethis run
- 96.9tok/s56.5Unknown hardware
- 96.8tok/s55.8Unknown hardware
- 96.7tok/s61.8Unknown hardware
- 96.6tok/s53.2Unknown hardware
Reproduce this run
toolllama.cpp b12192-9f0e5098a
modelqwen36-agent
context98,304
thinkingon
think limit4,096 tokens
samplingtemperature 0.699999988079071 · 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 "qwen36-agent" --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.