Benchmark result

qwen3:8b on AMD Radeon RX 9070 XT — 82.4 tok/s

Measured with Ollama 0.34.3 on September 23, 2026.

ShareRedditX

How this run compares

Reproduce this run

toolOllama 0.34.3
modelqwen3:8b (Q4_K_M)
context8,192
samplingtemperature 0.6 · top_p 0.95 · top_k 20 · repeat_penalty 1
clientv0.4.80+98

Set those in Ollama, then:

llm-benchmark benchmark --model "qwen3:8b" --tool "Ollama"

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.