Community performance summary

AMD Radeon Pro Vega II

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

17 signed runs · 0 verified · 5 testers
Median prompt483.53tok/s · 143.86 to 2289.14
Median generation48.18tok/s · 17.11 to 163.8
Configurations5Q4_K_M · Q4_K_S · Q8_0
Last observedOct 16GCN / 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 GPU9 groups
GaMS3 12B1 run→GLM 4 9B1 run→Qwen3 VL 2B3 runs→Qwen3 4B2 runs→V2.1 27B-A18B 27B-A18B1 run→Qwen3 Coder 30B-A3B2 runs→Qwen3.8 27B1 run→Qwen3.6 35B A3B3 runs→Gemma 4 26B A4B3 runs→
macOS

macOS 15.7.7 Sequoia · macOS 26.4 Tahoe · macOS 26.5.2 Tahoe · macOS 26.6 Tahoe · macOS 26.6.2 Tahoe · macOS 26.7 Tahoe · macOS 26.7.1 Tahoe

Recorded evidence

Inspect every run.

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

1 to 8 of 17 configurations

Verified benchmark Community submission
Model / quantHardwareArchitecturePromptGenerationSource
Qwen3 VL 2BQ8_0 · Dense · Signed ToshLLM app · amd-gpu · Metal · comparison
GCN / Vega
2289.14tok/s
163.8tok/s
ToshLLM communitySep 22, 2026
Qwen3 VL 2BQ8_0 · Dense · Signed ToshLLM app · amd-gpu · Metal · comparison
GCN / Vega
2220.1tok/s
161.48tok/s
ToshLLM communitySep 22, 2026
Qwen3 VL 2BQ8_0 · Dense · Signed ToshLLM app · amd-gpu · Metal · comparison
GCN / Vega
2275.38tok/s
96.88tok/s
simmc303Aug 26, 2026
Qwen3 4BQ4_K_M · Dense · Signed ToshLLM app · amd-gpu · Metal · comparison
GCN / Vega
1236.49tok/s
89.8tok/s
ToshLLM communitySep 22, 2026
Qwen3 4BQ4_K_M · Dense · Signed ToshLLM app · amd-gpu · Metal · comparison
GCN / Vega
1199.09tok/s
69.2tok/s
simmc303Aug 26, 2026
Qwen3 Coder 30B-A3BQ4_K_M · MoE · Signed ToshLLM app · amd-gpu · Metal · comparison
GCN / Vega
617.65tok/s
54.89tok/s
simmc303Aug 28, 2026
GLM 4 9BQ4_K_M · Dense · Signed ToshLLM app · amd-gpu · Metal · comparison
GCN / Vega
576.46tok/s
54.07tok/s
simmc303Sep 23, 2026
Qwen3 Coder 30B-A3BQ4_K_M · MoE · Signed ToshLLM app · amd-gpu · Metal · comparison
GCN / Vega
346.78tok/s
53.32tok/s
gervaiscoquil✓Jul 18, 2026
Gemma 4 26B-A4BUD-Q4_K_M · MoE · Signed ToshLLM app · amd-gpu · Metal · comparison
GCN / Vega
484.84tok/s
48.18tok/s
gervaiscoquil✓Jul 18, 2026
GCN / Vega
190.3tok/s
47.9tok/s
Mac Pro communityJul 16, 2026
Gemma 4 26B-A4BUD-Q4_K_M · MoE · Signed ToshLLM app · amd-gpu · Metal · comparison
GCN / Vega
483.53tok/s
47.63tok/s
ButterChickenJul 20, 2026
Qwen3.6 35B-A3BUD-Q4_K_S · MoE · Signed ToshLLM app · amd-gpu · Metal · comparison
GCN / Vega
280.34tok/s
42.24tok/s
gervaiscoquil✓Jul 18, 2026
Qwen3.6 35B A3BQ4_K_S · MoE · All experts in VRAM · comparison
GCN / Vega
284.7tok/s
42.1tok/s
Mac Pro communityJul 16, 2026
Qwen3.6 35B-A3BUD-Q4_K_S · MoE · Signed ToshLLM app · amd-gpu · Metal · comparison
GCN / Vega
432.42tok/s
41.97tok/s
ToshLLM communityAug 4, 2026
GaMS3 12BQ8_0 · Dense · Signed ToshLLM app · standard-auto · Metal · comparison
GCN / Vega
443.8tok/s
27.48tok/s
simmc303Oct 16, 2026
V2.1 27B-A18B 27B-A18BQ4_K_S · MoE · Signed ToshLLM app · amd-gpu · Metal · comparison
GCN / Vega
224.55tok/s
22.23tok/s
ToshLLM communitySep 22, 2026
Qwen3.8 27BUD-Q4_K_M · Dense · Signed ToshLLM app · amd-gpu · Metal · comparison
GCN / Vega
143.86tok/s
17.11tok/s
simmc303Aug 25, 2026