Benchmark result
qwen3-coder:30B on AMD Radeon Graphics — 103.0 tok/s
Measured with llama.cpp b11749-1d88cb05f on October 21, 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
qwen3-coder:30B on other hardware
- 240.2tok/s56.4NVIDIA GeForce RTX 5090 · Q4_K_M
- 103.0tok/s71.9Unknown hardwarethis run
- 84.5tok/s54.4Apple M5 Pro
- 82.9tok/s65.4Apple M5 Pro · Q4_K_M
- 56.5tok/s57.0Apple M5 Max
Reproduce this run
toolllama.cpp b11749-1d88cb05f
modelqwen3-coder:30B
context65,536
thinkingon
samplingtemperature 0.800000011920929 · top_p 0.949999988079071 · top_k 40 · min_p 0.05000000074505806 · presence_penalty 0 · repeat_penalty 1
clientv0.4.96+102
Set those in llama.cpp, then:
llm-benchmark benchmark --model "qwen3-coder:30B" --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.