Community performance summary

RX 6900 XT + Vega

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

6 signed runs · 0 verified · 1 tester
Median prompt1680.32tok/s · 763.1 to 6224.67
Median generation62.17tok/s · 30.98 to 206.18
Configurations1Q4_K_M
Last observedAug 10RDNA 2 + GCN / 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.

Recorded evidence

Inspect every run.

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

1 to 8 of 6 configurations

Verified benchmark Community submission
Model / quantHardwareArchitecturePromptGenerationSource
Llama 3.2 1BQ4_K_M · Dense · Signed ToshLLM app · amd-gpu · Metal · comparison
RDNA 2 + GCN / Vega
6224.67tok/s
206.18tok/s
ToshLLM communityAug 10, 2026
Llama 3.2 1BQ4_K_M · Dense · Signed ToshLLM app · amd-gpu · Metal · comparison
RDNA 2 + GCN / Vega
3047.25tok/s
113.04tok/s
ToshLLM communityAug 10, 2026
Qwen3 4BQ4_K_M · Dense · Signed ToshLLM app · amd-gpu · Metal · comparison
RDNA 2 + GCN / Vega
2143.87tok/s
81.82tok/s
ToshLLM communityAug 10, 2026
Qwen3 4BQ4_K_M · Dense · Signed ToshLLM app · amd-gpu · Metal · comparison
RDNA 2 + GCN / Vega
1216.76tok/s
42.52tok/s
ToshLLM communityAug 10, 2026
Qwen3 Coder 30B-A3BQ4_K_M · MoE · Signed ToshLLM app · amd-gpu · Metal · comparison
RDNA 2 + GCN / Vega
807.25tok/s
33.2tok/s
ToshLLM communityAug 10, 2026
Gemma 4 12BQ4_K_M · Dense · Signed ToshLLM app · amd-gpu · Metal · comparison
RDNA 2 + GCN / Vega
763.1tok/s
30.98tok/s
ToshLLM communityAug 10, 2026