FLUX 3 Image Klein is a state-of-the-art open-weight image generation and editing model Nemotron 3.5 Super early-access reasoning model for agentic, coding, tool-calling, and multimodal workloads Preview of StepFun's Step 5: ~600B-total / 27B-active multimodal MoE (text and image in, text out) with 1M context, reasoning and tool calling, served in NVFP4. A tabular in-context learning model for classification and regression over structured tables. Estimate 3D human body pose and skeleton from video input. 552B MoE, 8B active params with native multimodal support and lower API cost using smaller KV cache A relational foundation model for prediction over structured, multi-table data. 753B-parameter text MoE with DeepSeek-style sparse attention, native FP8 weights, reasoning and tool calling. Multimodal 320B-total / 18B-active MoE with hybrid KDA and sparse MLA attention, native FP8 weights, reasoning and tool calling. Cutting-edge vision-language model excelling in retrieving text and metadata from images. Multilingual ASR across 25 European languages with punctuation, capitalization, and word timestamps Wan2.2-Animate-2 is a novel end-to-end character animation framework Fastest 30B A3B MoE model with leading domain accuracy for specialized agentic tasks Muse Glimmer 30B is a multimodal reasoning model accepting text and images, with native tool-calling and separate reasoning output. Translation model in 37 languages with few-shots example prompts capability. NVIDIA-Ising-Calibration-1.5 is a dense multimodal vision-language model built on Gemma 4 31B. It analyzes quantum computing calibration experiment plots and generates structured technical text. Upscale encoded or ST 2110 video to higher resolutions with NVIDIA Video Super Resolution. 1B embedding model for semantic search, retrieval, and RAG applications. Efficient 33B MoE for local, long-horizon agentic coding and terminal tasks An image edit model specialized for Omniverse synthetic to photographic solder-light style captured at NVIDIA PCB inspection stations Nemotron OCR v2 is a state-of-the-art multilingual text recognition model designed for robust end-to-end optical character recognition (OCR) on complex real-world images. Diffusion-based 26B parameter LLM enabling parallel token generation for real-time text apps Open, efficient hybrid Mamba-Transformer MoE with 1M context, excelling in agentic reasoning, coding, planning, tool calling, and more Natural and expressive voices in 23 languages. For voice agents and brand ambassadors. Image Generation Large Language Models Structured Data In-Context Learning nvidia ai for media

Vision Language Model KV Cache Optimization Causal Encoder Decoder Structured Data Relational Models Large Language Models Large Language Models text and table extraction document parsing supported language - english character animation Long-running agents Large Language Models neural machine translation Text Translation Quantum Computing Vision Language Model video upscaling nvidia ai for media video super resolution Nemotron Retriever Agentic Retrieval Text-to-Embedding Retrieval Augmented Generation Synthetic Data Generation Image Generation Table Extraction Optical Character Recognition Speech Generation