Flow - 4x faster than typing
Mini toolsNemotron models
Browse all models from this model family.
ID
Model
Company
Type
Primary task
Open source
-
By NVIDIANemotron-3-Embed-1B-NVFP4 is a hardware accelerated, 8-bit quantized variant of NVIDIA's Nemotron-3-Embed-1B-BF16 text embedding model. It shares the same embedding space and produces dense multilingual text embeddings for retrieval and RAG, optimized for high throughput, massive scale inference on Blackwell GPUs.NewTextReleased 1d ago
-
By NVIDIANemotron-3-Embed-1B-BF16 is NVIDIA's high efficiency multilingual text embedding model, pruned and distilled from the 8B flagship. It produces dense vector embeddings for retrieval, semantic search, and RAG across 34 languages, built for low latency, cost sensitive production serving.NewTextReleased 1d ago
-
By NVIDIANemotron-3-Embed-8B-BF16 is NVIDIA's flagship multilingual text embedding model, built on a Ministral-3-8B based encoder. It generates dense vector representations for retrieval, semantic search, and RAG across 34 languages, and ranks first overall on the RTEB benchmark for embedding quality and agentic retrieval performance.NewTextReleased 1d ago
-
By NVIDIANemotron-Labs-Audex-30B-A3B is a unified audio-text large language model from NVIDIA built on a 30B parameter Mixture-of-Experts backbone with 3B active parameters. It extends a strong text-only reasoning model with an audio encoder and discrete audio token vocabulary, enabling audio understanding, speech recognition, speech translation, text-to-speech, audio generation, and speech-to-speech interaction while preserving the backbone's reasoning, knowledge, and long-context abilities.NewMultimodalReleased 1mo ago
-
By NVIDIANVIDIA Nemotron-3-Ultra-550B-A55B-NVFP4 is NVIDIA’s open 550B total, 55B active frontier reasoning LLM for agentic workflows, long-context analysis, tool use, code, math, and science.NewTextReleased 1mo ago
-
By NVIDIANemotron-Labs-Diffusion-14B is NVIDIA’s open 14B text-generation language model that supports autoregressive decoding, diffusion-style parallel decoding, and self-speculative decoding.NewTextReleased 1mo ago
-
By NVIDIANemotron 3 Nano Omni is NVIDIA’s open multimodal reasoning model for agentic systems. It unifies text, image, video, and audio in a single efficient 30B-A3B hybrid MoE model, built to replace fragmented vision-language-audio stacks with one shared perception-and-context model for multimodal agents.NewMultimodalReleased 2mo ago
-
By NVIDIANemotron-Cascade 2 is an open 30B Mixture-of-Experts model (3B activated) trained with Cascade RL and multi-domain on-policy distillation for strong reasoning and agentic capabilitieTextReleased 4mo ago
-
By NVIDIANemotron 3.5 ASR is a 600M parameter multilingual streaming automatic speech recognition model supporting 40 language locales in a single checkpoint. Built on a cache-aware FastConformer-RNNT architecture, it maintains encoder context across audio chunks for accurate real-time transcription with native punctuation and capitalization output.AudioReleased 4mo ago
-
By NVIDIANemotron 3 Super is NVIDIA’s open Mixture-of-Experts hybrid Mamba-Transformer model designed for high-throughput agentic workloads and long-context reasoning.TextReleased 4mo ago
-
By NVIDIANemotron-3-Nano-30B-A3B is NVIDIA’s 30B-class LLM trained from scratch as a unified model for both reasoning and standard chat tasks. It can be configured to produce a reasoning trace before the final answer, letting developers switch reasoning behavior through the chat template.TextReleased 6mo ago
-
By NVIDIANVIDIA Nemotron 3 is an open family of models (Nano, Super, and Ultra) built on a hybrid Mamba-Transformer Mixture-of-Experts architecture for efficient agentic AI, reasoning, and conversational workloads, with context windows up to 1M tokens.TextReleased 7mo ago
-
TextReleased 7mo ago
-
By NVIDIALlama-3.3-Nemotron-Super-49B-v1.5 is NVIDIA’s upgraded 49B reasoning model derived from Meta Llama-3.3-70B-Instruct. It is post-trained for reasoning, chat preference alignment, and agentic tasks such as RAG and tool use.TextReleased 9mo ago
-
By NVIDIANemotron-4 is NVIDIA’s open-weight LLM family (from compact to 340B) built for high-quality reasoning, coding, and synthetic data generation. It supports tool/function calling, JSON outputs, and long-context variants, and is production-ready via NVIDIA NIM/TensorRT-LLM for fast, scalable deployment.TextReleased 9mo ago
-
TextReleased 10mo ago
-
By NVIDIANemotron-Nano 2 is NVIDIA’s ultra-compact LLM tuned for on-device and edge deployment. It delivers fast instruction following, coding help, and reasoning with low memory use, supports tool/function calling and structured (JSON) outputs, and runs efficiently on Jetson, RTX laptops/desktops, and server GPUs.TextReleased 10mo ago
-
By NVIDIALlama 3.1 Nemotron Ultra is an NVIDIA-optimized deployment of Meta’s Llama 3.1, packaged for high-throughput production. It delivers strong reasoning and coding, long-context support (≈128K), tool/function calling, and JSON mode—served as a fast, scalable NIM for apps and agents.TextReleased 1y ago
-
By NVIDIACosmos Nemotron VLM is NVIDIA’s multimodal model that fuses Cosmos world-model perception with Nemotron language reasoning. It understands images and video alongside text, performs step-by-step visual reasoning, and supports tool/function calling and JSON outputs—optimized for fast, scalable deployment via TensorRT-LLM and NIM.TextReleased 1y ago
