Nvidia
California, 美国NVIDIA’s open 120B-parameter vision-language MoE model with 12B active parameters, built for reasoning, coding, tool use, document understanding and video understanding.
Nemotron-Lightning-3.5-30B-A3B is a 30B-parameter Mixture-of-Experts language model (3B active) from NVIDIA's Nemotron-H family, built on a hybrid Mamba-Transformer architecture for efficient long-context inference. Like other models in the family, it responds to queries by first generating a reasoning trace and then concluding with a final response, with reasoning behavior configurable through a flag in the chat template. It includes a multi-token prediction (MTP) speculative decoding head for low-latency serving.
Nemotron 3.5 Lightning is NVIDIA's open 30B-parameter mixture-of-experts reasoning model with 3B active parameters, a hybrid Mamba-Transformer architecture, and support for high-throughput agentic workloads.
Nemotron 3 Embed 1B BF16 is a multilingual text embedding model for retrieval, semantic search, and RAG workflows.
Nemotron 3 Embed 8B BF16 is a multilingual text embedding model for retrieval, semantic search, and RAG workflows.
NVIDIA Nemotron 3 Ultra is NVIDIA's strongest open-weights reasoning model, positioned near GPT-5.4 Mini (xhigh) and ahead of DeepSeek V4-Flash and Qwen3.5-397B-A17B.
NVIDIA Nemotron 3.5 Content Safety is a compact 4B-parameter multimodal guardrail model from NVIDIA, fine-tuned from Google Gemma-3-4B. It moderates both inputs to and responses from LLMs and VLMs, accepting text and image input and returning text output: a safe/unsafe classification for the user prompt and the response, safety category labels, and an optional reasoning trace. It covers 12 languages with a context window of up to 128K tokens. Suited for prompt and response moderation, content classification, safety pipelines, and enterprise AI guardrails with policy enforcement, and includes a togglable reasoning mode. Part of the NVIDIA Nemotron family of open models for agentic AI.
Nemotron-3-Ultra-550B-A55B-NVFP4 is a frontier-scale large language model (LLM) trained by NVIDIA, designed to deliver strong agentic, reasoning, and conversational capabilities. It is optimized for the most demanding workloads, including complex multi-step agents, long-context analysis, and high-accuracy reasoning over code, math, and science. The model employs a hybrid Latent Mixture-of-Experts (LatentMoE) architecture, utilizing interleaved Mamba-2 and MoE layers, along with select Attention layers. Like the Super model, the Ultra model incorporates Multi-Token Prediction (MTP) layers for faster text generation and improved quality, and it is trained using an NVFP4 pre-training recipe to maximize compute efficiency. The model has 55B active parameters and 550B parameters in total.
Nemotron 3 Ultra is NVIDIA's largest Nemotron 3 model, a 550B total-parameter hybrid Mamba-Transformer mixture-of-experts model with up to 55B active parameters per token and up to 1M tokens of context, built for coding, research, and long-running agentic workflows.
Cosmos3 Super Reasoner is an NVIDIA reasoning model for physical-AI and agent workflows, designed to answer from long-context text, image, and video inputs.
Nemotron 3 Nano Omni 30B A3B Reasoning is a 30B-A3B open multimodal model designed to function as a perception and context sub-agent in enterprise agent systems.
Nemotron Cascade 2 30B A3B is NVIDIA's open 30B mixture-of-experts reasoning model with 3B active parameters, supporting thinking and non-thinking modes for agentic, mathematical, and coding work.
Nemotron 3 VoiceChat is NVIDIA's 12B full-duplex speech-to-speech model, hosted by Crusoe Serverless Inference with audio and text interaction.
Nemotron 3 Super is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications.
Nemotron Nano 3 30B A3B is a small language MoE model with highest compute efficiency and accuracy for developers to build specialized agentic AI systems.
Llama 3.1 Nemotron Ultra 253B v1 is a 49B-parameter, English-centric reasoning/chat model derived from Meta's Llama-3.3-70B-Instruct with a 128K context.
Llama 3.1 Nemotron Nano 8B V1 is a 49B-parameter, English-centric reasoning/chat model derived from Meta's Llama-3.3-70B-Instruct with a 128K context.
Llama 3.3 Nemotron Super 49B v1 is a 49B-parameter, English-centric reasoning/chat model derived from Meta's Llama-3.3-70B-Instruct with a 128K context.
Llama 3.1 Nemotron 70B Instruct is a 49B-parameter, English-centric reasoning/chat model derived from Meta's Llama-3.3-70B-Instruct with a 128K context.
NVIDIA Riva text-to-speech model for American English.
Nemotron 3 Nano Omni is a DigitalOcean-hosted NVIDIA model exposed through DigitalOcean Inference.
Nemotron Nano 12B V2 VL is a multimodal model from NVIDIA that can work across text and visual inputs. It is a smaller-footprint option suited to tighter latency or deployment budgets.
Nemotron Nano 9B V2 is a large language model (LLM) trained from scratch by NVIDIA, and designed as a unified model for both reasoning and non-reasoning tasks.
Llama 3.1 Nemotron Nano 4B V1.1 is a 49B-parameter, English-centric reasoning/chat model derived from Meta's Llama-3.3-70B-Instruct with a 128K context.
Llama 3.3 Nemotron Super 49B V1.5 is a 49B-parameter, English-centric reasoning/chat model derived from Meta's Llama-3.3-70B-Instruct with a 128K context.
Nemotron Nano 12B V2 is a large language model (LLM) trained from scratch by NVIDIA, and designed as a unified model for both reasoning and non-reasoning tasks.
Openreasoning Nemotron 1.5B is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications.
Openreasoning Nemotron 14B is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications.
Openreasoning Nemotron 32B is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications.
Openreasoning Nemotron 7B is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications.