NVIDIA Nemotron 3 Embed Ranks #1 Overall on RTEB, Advancing Agentic Retrieval
NVIDIA Nemotron 3 Embed Ranks #1 Overall on RTEB, Advancing Agentic Retrieval
NVIDIA Nemotron 3 Embed 在 RTEB 榜单中排名第一，推动智能体检索技术发展
Retrieval is critical in multi-step agentic workflows where poor retrieval can cause agents to fetch irrelevant context, re-query, waste token budget, and carry noise into later reasoning steps. Today, we are releasing NVIDIA Nemotron 3 Embed, a collection of open and commercially available embedding models designed to improve retrieval quality while giving developers practical deployment options for production-scale RAG, agentic retrieval, code retrieval, and agent memory.
在多步智能体（Agentic）工作流中，检索至关重要。糟糕的检索会导致智能体获取无关上下文、重复查询、浪费 Token 预算，并将噪声引入后续的推理步骤。今天，我们发布了 NVIDIA Nemotron 3 Embed，这是一系列开源且可商用的嵌入（Embedding）模型，旨在提升检索质量，同时为开发者提供适用于生产级 RAG、智能体检索、代码检索和智能体记忆的实用部署方案。
The collection includes three open models that achieve state-of-the-art retrieval across the accuracy-efficiency curve, led by an 8B model that tops the RTEB leaderboard and efficient 1B variants built for production-scale deployment:
该系列包含三个开源模型，在准确性与效率之间实现了业界领先的检索表现。其中，8B 模型登顶 RTEB 榜单，而高效的 1B 变体则专为生产级部署而打造：
| Model | Role | Best for |
|---|---|---|
| Nemotron-3-Embed-8B-BF16 | Flagship Quality Anchor | Precision-critical retrieval and high-stakes enterprise RAG |
| Nemotron-3-Embed-1B-BF16 | High-Efficiency Standard | Cost- and latency-sensitive production serving |
| Nemotron-3-Embed-1B-NVFP4 | Hardware-Accelerated Variant | Ultra-high-throughput and massive-scale infrastructure |
| 模型 | 角色 | 适用场景 |
|---|---|---|
| Nemotron-3-Embed-8B-BF16 | 旗舰质量标杆 | 精度要求极高的检索及高风险企业级 RAG |
| Nemotron-3-Embed-1B-BF16 | 高效标准版 | 对成本和延迟敏感的生产环境服务 |
| Nemotron-3-Embed-1B-NVFP4 | 硬件加速变体 | 超高吞吐量及大规模基础设施 |
Key Features
核心功能
Beyond the RTEB result, Nemotron 3 Embed introduces a production-ready feature set for enterprise retrieval deployments: 除了在 RTEB 榜单上的表现，Nemotron 3 Embed 还为企业级检索部署引入了一套生产就绪的功能集：
- Open Weights, Datasets, and Recipes: Gives teams control to inspect, tune, fine-tune, and deploy retrieval models on their own infrastructure. 开放权重、数据集和配方： 使团队能够自主检查、调整、微调并在自有基础设施上部署检索模型。
- 32k Context Window: Supports retrieval over long documents, large code contexts, and multi-turn agent histories while reducing truncation. 32k 上下文窗口： 支持对长文档、大型代码上下文和多轮智能体历史记录进行检索，同时减少截断。
- Multilingual & Code Retrieval: Supports retrieval across global enterprise data, technical documentation and multi-file code repositories. 多语言与代码检索： 支持跨全球企业数据、技术文档和多文件代码库的检索。
- NVIDIA NVFP4 Efficiency: Provides a Blackwell-optimized 4-bit deployment path for high-throughput retrieval with a smaller memory footprint. NVIDIA NVFP4 效率： 提供针对 Blackwell 架构优化的 4-bit 部署路径，以更小的内存占用实现高吞吐量检索。
- Fine-Tuning and Distillation Recipes: NVIDIA NeMo AutoModel recipes support domain adaptation and model compression for teams adapting retrieval models to their own data. 微调与蒸馏配方： NVIDIA NeMo AutoModel 配方支持领域自适应和模型压缩，帮助团队将检索模型适配到自有数据。
- Day-0 Ecosystem Integration: Available immediately on Hugging Face, deployable as NVIDIA NIM microservice, supported by vLLM, and accessible through leading AI Cloud and inference partners. 零日生态集成： 即刻在 Hugging Face 上可用，可作为 NVIDIA NIM 微服务部署，支持 vLLM，并可通过领先的 AI 云和推理合作伙伴访问。
Evaluation: Retrieval Quality, Agentic Efficiency, and Deployment Tradeoffs
评估：检索质量、智能体效率与部署权衡
We evaluate Nemotron 3 Embed across three dimensions: retrieval quality, downstream agentic efficiency, and deployment tradeoffs. The 8B model establishes the model collection’s quality ceiling, while the 1B BF16 and NVFP4 variants bring the same retrieval-focused design to lower-cost and higher-throughput deployment settings.
我们从三个维度评估 Nemotron 3 Embed：检索质量、下游智能体效率和部署权衡。8B 模型确立了该系列模型的质量上限，而 1B BF16 和 NVFP4 变体则将同样以检索为中心的设计带入了更低成本、更高吞吐量的部署场景。
RTEB Leadership and Strong Gains Across Retrieval Benchmarks
RTEB 领先地位与检索基准测试的显著提升
We first evaluated the models on RTEB, where Nemotron-3-Embed-8B-BF16 ranks #1. We also tested these models across ViDoRe V3 Text, and MMTEB Retrieval and LongEmbed using average NDCG@10.
我们首先在 RTEB 上评估了这些模型，Nemotron-3-Embed-8B-BF16 排名第一。我们还使用平均 NDCG@10 在 ViDoRe V3 Text、MMTEB Retrieval 和 LongEmbed 上对这些模型进行了测试。
Nemotron-3-Embed-8B-BF16 ranks #1 on RTEB, scoring 78.5% on RTEB and 75.5% on MMTEB Retrieval. Nemotron-3-Embed-1B-BF16 brings much of the 8B model’s retrieval quality into a smaller deployment footprint. It scores 72.4% on RTEB, reducing error rate by 27% over its 1B predecessor (llama-nemotron-embed-vl-1b-v2), and scores 71.0% on MMTEB Retrieval, reducing error rate by 28%.
Nemotron-3-Embed-8B-BF16 在 RTEB 上排名第一，得分 78.5%，在 MMTEB Retrieval 上得分 75.5%。Nemotron-3-Embed-1B-BF16 将 8B 模型的大部分检索质量带入了更小的部署空间。它在 RTEB 上得分 72.4%，较其 1B 前代产品（llama-nemotron-embed-vl-1b-v2）错误率降低了 27%，在 MMTEB Retrieval 上得分 71.0%，错误率降低了 28%。
Why Better Retrieval Matters for Agents
为什么更好的检索对智能体至关重要
To evaluate retrieval in an agentic setting, we use a search agent powered by Nemotron 3 Ultra and vary the embedding model used by the retrieval system. Better retrieval can return relevant evidence earlier, helping the agent avoid repeated searches, unnecessary reasoning turns, and extra context inspection.
为了在智能体环境中评估检索效果，我们使用由 Nemotron 3 Ultra 驱动的搜索智能体，并改变检索系统所使用的嵌入模型。更好的检索能够更早地返回相关证据，帮助智能体避免重复搜索、不必要的推理轮次以及额外的上下文检查。
Figure 3 shows that stronger retrieval reduces downstream agentic token cost. More accurate retrievers return relevant evidence earlier, which helps agents complete tasks with fewer repeated searches and fewer reasoning turns. In these evaluations, the Nemotron 3 Embed models improve the agentic retrieval frontier, with the 8B model delivering both the highest average retrieval accuracy and the lowest estimated downstream token cost across ViDoRe V3, BRIGHT, and BrowseComp-Plus.
图 3 显示，更强的检索能力降低了下游智能体的 Token 成本。更准确的检索器能更早返回相关证据，从而帮助智能体以更少的重复搜索和推理轮次完成任务。在这些评估中，Nemotron 3 Embed 模型提升了智能体检索的边界，其中 8B 模型在 ViDoRe V3、BRIGHT 和 BrowseComp-Plus 测试中，既提供了最高的平均检索准确率，也实现了最低的预估下游 Token 成本。
Scaling Retrieval with NVFP4 on Blackwell
在 Blackwell 上利用 NVFP4 扩展检索
For high-throughput deployments, teams often choose smaller embedding models to meet latency and cost targets. Nemotron-3-Embed-1B-NVFP4 is designed to narrow the gap between serving efficiency and retrieval.
对于高吞吐量部署，团队通常会选择较小的嵌入模型以满足延迟和成本目标。Nemotron-3-Embed-1B-NVFP4 旨在缩小服务效率与检索质量之间的差距。