Benchmark result

Ornith-35B-A3B on AMD Radeon RX 7800 XT — 89.7 tok/s

Measured with ornith-35b-a3b-think manual on November 16, 2026.

ShareRedditX

Support llm-bench.io

llm-bench.io is free. If it helped you, I'd appreciate your support to cover the infrastructure costs.

Support me on Ko-fiBuy me a coffee

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.

Reproduce this run

toolornith-35b-a3b-think manual
modelOrnith-35B-A3B
context8,192
thinkingon
clientv0.4.99+102

Set those in ornith-35b-a3b-think, then:

llm-benchmark benchmark --model "Ornith-35B-A3B" --tool "ornith-35b-a3b-think"

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.