Gemma 4 26B-A4B

UD-IQ3_S running on AMD Radeon RX Ellesmere Prototype with macOS 15.7.3 Sequoia

PERFORMANCE / TOKENS PER SECONDUnknown

Prompt processing116.33tok/s

#117 in index

Token generation7.48tok/s

#123 in index

Run fingerprint

Enough context to make the number useful.

ModelGemma 4 26B-A4BMoE architecture

QuantizationUD-IQ3_SExact submitted variant

GPUAMD Radeon RX Ellesmere Prototype8 GB VRAM

GPU architectureUnknownMetal backend

Test context512prompt tokens

Model familyMoESigned ToshLLM app · ncmoe 3 · amd-gpu · Metal

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.