Benchmark result
Qwen3.8-Flash-Next-Avalon-IQ3KS-00001-of-00033 on Advanced Micro Devices, Inc. [AMD/ATI] Device 7550 — 43.7 tok/s
Measured with llama.cpp b12351-dc6169adf on November 14, 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
8th fastest of 22 runs of this model on Advanced Micro Devices, Inc. [AMD/ATI] Device 7550 · median 41.4 tok/s.
- 145.1tok/s79.7llama.cpp
- 245.0tok/s84.0llama.cpp
- 344.9tok/s84.8llama.cpp
- 444.8tok/s87.8llama.cpp
- 544.4tok/s64.5llama.cpp
- 644.3tok/s88.5llama.cpp
- 744.0tok/s86.6llama.cpp
- 843.7tok/s82.7llama.cppthis run
- 942.7tok/s82.3llama.cpp
- 1041.7tok/s82.7llama.cpp
- 1141.5tok/s83.3llama.cpp
- 1241.2tok/s44.4llama.cpp
- 1340.8tok/s84.1llama.cpp
- 1440.1tok/s88.5llama.cpp
- 1539.1tok/s84.0llama.cpp
- 1638.2tok/s79.7llama.cpp
- 1737.5tok/s87.9llama.cpp
- 1837.5tok/s85.1llama.cpp
- 1937.3tok/s81.2llama.cpp
- 2036.7tok/s83.4llama.cpp
- 2136.3tok/s84.6llama.cpp
- 2236.3tok/s85.4llama.cpp
Qwen3.8-Flash-Next-Avalon-IQ3KS-00001-of-00033 on other hardware
Reproduce this run
Set those in llama.cpp, then:
llm-benchmark benchmark --model "Qwen3.8-Flash-Next-Avalon-IQ3KS-00001-of-00033" --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.