Community performance summary

AMD Radeon Pro W5700X ×2

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

4 signed runs · 0 verified · 2 testers
Median prompt703.91tok/s · 127.25 to 791.23
Median generation48.01tok/s · 14.25 to 59.61
Configurations4IQ4 · Q4_K_M · Q5_K_M
Last observedOct 12RDNA 1

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 GPU4 groups
Ornith 1.5 35B1 run→Ornith 1.0 35B1 run→Qwen3.6 35B-A3B1 run→Qwen3.6 27B1 run→
macOS

macOS 26.5 Tahoe · macOS 26.6.1 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 4 configurations

Verified benchmark Community submission
Model / quantHardwareArchitecturePromptGenerationSource
Ornith 1.5 35BQ5_K_M · MoE · Signed ToshLLM app · amd-gpu · Metal · comparison
RDNA 1
791.23tok/s
59.61tok/s
HypabolicOct 12, 2026
Ornith 1.0 35BQ4_K_M · MoE · Signed ToshLLM app · amd-gpu · Metal · comparison
RDNA 1
703.09tok/s
49.49tok/s
HypabolicAug 12, 2026
Qwen3.6 35B-A3BUD-Q4_K_M · MoE · Signed ToshLLM app · amd-gpu · Metal · comparison
RDNA 1
704.73tok/s
46.53tok/s
HypabolicAug 12, 2026
Qwen3.6 27BIQ4 · Dense · Signed ToshLLM app · amd-gpu · Metal · comparison
RDNA 1
127.25tok/s
14.25tok/s
malzzzJul 23, 2026