Build More Accurate and Efficient AI Agents with the New NVIDIA Llama Nemotron Super v1.5

The NVIDIA Nemotron family builds on the strongest open models in the ecosystem by enhancing them with greater accuracy, efficiency, and transparency using NVIDIA open synthetic datasets, advanced techniques, and tools. Today, we’re introducing NVIDIA Llama Nemotron Super v1.5, which brings significant improvements across core reasoning and agentic tasks like math, science, coding, function calling, instruction following, and chat, while maintaining strong throughput and compute efficiency.

Built for reasoning and agentic workloads

Llama Nemotron Super v1.5 builds on the same efficient reasoning foundation as Llama Nemotron Ultra. However, the model has been refined through post-training using a new dataset focused specifically on high-signal reasoning tasks. Across a wide range of benchmarks, Llama Nemotron Super v1.5 outperforms other open models in its weight class, particularly in tasks that require multi-step reasoning and structured tool use.

Figure 1. Llama Nemotron Super v1.5 delivers the highest accuracy for reasoning and agentic tasks.

To boost throughput and deployment efficiency, pruning techniques such as neural architecture search were applied. Higher throughput means the model can reason faster and explore more complex problem spaces within the same compute and time budget—delivering stronger reasoning at lower inference costs. It also runs on a single GPU, further reducing compute overhead.

Figure 2. Llama Nemotron Super v1.5 provides the highest accuracy and throughput for agentic tasks, lowering the cost of inference.

Try the model now

Experience Llama Nemotron Super v1.5 at build.nvidia.com, or download the model directly from Hugging Face.

Related resources

• GTC session: Building Scalable Data Flywheels for Continuously Improving AI Agents
• GTC session: Building Future-Ready AI With Agents and Data Flywheels: Insights From NVIDIA’s Enterprise Deployments
• NGC Containers: Llama-3.1-Nemotron-70B-Instruct
• SDK: Llama3 70B Instruct NIM
• SDK: Llama3 8B Instruct NIM
• Webinar: Accelerating Contact Center AI Workflows with NVIDIA AI Enterprise

---

Source attribution: Originally published at NVIDIA Developer Blog on July 25, 2025.

FAQ

What are the key enhancements in NVIDIA Llama Nemotron Super v1.5?

NVIDIA Llama Nemotron Super v1.5 features enhanced accuracy, efficiency, and transparency with augmented reasoning and agentic task capabilities, improved using NVIDIA open synthetic datasets and advanced techniques.

How does Llama Nemotron Super v1.5 improve reasoning tasks?

By utilizing refined post-training datasets focused on high-signal reasoning tasks, Llama Nemotron Super v1.5 excels in complex multi-step reasoning and structured tool use, outclassing other models in its category.

How is throughput enhanced in Llama Nemotron Super v1.5?

Pruning techniques like neural architecture search boost the throughput, enabling faster reasoning capabilities and exploration of complex problem spaces with reduced compute costs, by operating on a single GPU.

Where can I experience Llama Nemotron Super v1.5?

You can try Llama Nemotron Super v1.5 at build.nvidia.com or download it directly from Hugging Face. ##Crypto Market's Take At Crypto Market, we recognize the significant advancements made by NVIDIA's Llama Nemotron Super v1.5 model. Leveraging advanced AI reasoning, similar to what powers our AI-driven trading bots, enhances complex decision-making in cryptocurrency trading. Our AI tools offer sophisticated analysis and automation, complementing NVIDIA's innovative approaches to AI development. Explore more about our AI Analysts and Trading Bots by visiting our AI Agents section. ##More to Read:
• AI-Driven Crypto Trading Tools: Transforming Market Strategies
• Agentic AI and Its Impact on the Crypto Sector
• Leveraging AI in Cryptocurrency