Gemma 4 26B-A4B

UD-Q6_K running on AMD Radeon Pro 5600M with macOS 26.6.2 Tahoe

PERFORMANCE / TOKENS PER SECONDRDNA 1

Prompt processing113.05tok/s

#259 in index

Token generation10.19tok/s

#266 in index

Run fingerprint

Enough context to make the number useful.

ModelGemma 4 26B-A4BMoE architecture

QuantizationUD-Q6_KExact submitted variant

GPUAMD Radeon Pro 5600M7.98 GB VRAM

GPU architectureRDNA 1Metal backend

Test context512prompt tokens

Model familyMoESigned ToshLLM app · ncmoe 40 · amd-gpu · Metal

ToshLLM versionv0.86.4App version used to capture and sign this run

Prompt processing and generation are reported separately so memory bandwidth and decode behavior stay visible.

Quantization, GPU family, VRAM and context travel with the result instead of being hidden behind one score.

This result remains explicitly labeled as a community submission.