Llama-4-Maverick-17B-128E-Instruct config.json

Same text_config width as Scout (hidden 5120, 48 layers, 40/8 heads, head_dim 128) but 128 local experts, 1 per token, no rope_scaling, max_position_embeddings 1048576. | SOURCE AUDIT 2026-10-03: kind was 'spec-sheet' but the URL is a raw Hugging Face config.json served as text/plain, not a vendor spec sheet; set to 'repo'. The schema field it fills (layer/window/head counts) is machine-readable model metadata and should be read as such.