Llm
2 prompts in this category
Optimized prompts for NVIDIA Nemotron 3 Ultra — the first open-weight 550B hybrid Mamba-MoE model. 55B active parameters, 1M context window, 89.1 MMLU. Datacenter-scale agentic reasoning.
Optimized prompts for Google Gemma 4 12B — the encoder-free any-to-any multimodal model. Handles text, image, audio, and video with 256K context. Apache 2.0 open weights. Laptop-class deployment.