GPT OSS 20B

MXFP4 running on AMD Radeon Pro W5700X

PERFORMANCE / TOKENS PER SECONDRDNA 1

Prompt processing705.92tok/s

#14 in index

Token generation73.3tok/s

#12 in index

Run fingerprint

Enough context to make the number useful.

ModelGPT OSS 20BMoE architecture

QuantizationMXFP4Exact submitted variant

GPUAMD Radeon Pro W5700X15.98 GB VRAM

GPU architectureRDNA 1Metal backend

Test context512prompt tokens

Model familyMoESigned ToshLLM app · 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.