Community performance summary

VII + RX Ellesmere Prototype

2 local models measured on this GPU configuration. Compare quantizations, configurations and the full observed performance range.

2 signed runs · 0 verified · 1 tester
Median prompt425.97tok/s · 26.84 to 825.1
Median generation32.8tok/s · 10.63 to 54.97
Configurations2UD-IQ4_XS · UD-Q4_K_XL
Last observedJul 21GCN / Vega

What the results show

One clear view.
Every recorded run.

Measurements from comparable hardware and models are presented together. Medians show typical performance, while every signed run remains available below for inspection and sharing.

Models on this GPU2 groups
Gemma 4 E4B1 runQwen3.6 35B-A3B1 run
macOS

macOS 15.7.3 Sequoia

Recorded evidence

Inspect every run.

Open any result to review its exact hardware, model configuration, measurements and submission details.

1 to 8 of 2 configurations

Verified benchmark Community submission
Model / quantHardwareArchitecturePromptGenerationSource
Gemma 4 E4BUD-Q4_K_XL · Dense · Signed ToshLLM app · amd-gpu · Metal · comparison
GCN / Vega
825.1tok/s
54.97tok/s
1252erJul 21, 2026
Qwen3.6 35B-A3BUD-IQ4_XS · MoE · Signed ToshLLM app · amd-gpu · Metal · comparison
GCN / Vega
26.84tok/s
10.63tok/s
1252erJul 21, 2026