---
title: "Nemotron 3 Ultra (NVIDIA) - Financial AI Performance - FinanceBenchmark"
description: "Independent evaluation and benchmark scores for Nemotron 3 Ultra on financial reasoning."
canonical: "https://financebenchmark.ai/models/nemotron-3-ultra"
type: "text/markdown"
---
# Nemotron 3 Ultra

> Developed by **NVIDIA** · Released 2026-06-04 · License: open-weight

## Table of Contents
- [1. Model Profile & Summary](#profile)
- [2. FB Composite Score](#composite-score)
- [3. All Evaluated Benchmarks](#benchmarks)
- [4. Certification & Verification](#verification)

## 1. Model Profile & Summary {#profile}

- **Organization:** NVIDIA
- **Release Date:** 2026-06-04
- **License:** open-weight (Open-Weight)
- **Context Window:** 1M
- **FB Certified:** Standard Attribution

## 2. FB Composite Score {#composite-score}

**Composite FB Score: 46.3 / 100**

- Benchmarks Evaluated: **16 of 42**

## 3. All Evaluated Benchmarks {#benchmarks}

| Benchmark | Category | Overall Score | Eval Date | Source |
|:----------|:---------|:-------------:|:---------:|:------:|
| [VALS Index](https://financebenchmark.ai/benchmarks/vals-index.md) | verification, numerical, agentTasks | **27.4%** | 2026-06 | [Paper](https://vals.ai) |
| [Code Migration](https://financebenchmark.ai/benchmarks/code-migration.md) | agentTasks | **4.9%** | 2026-06 | [Paper](https://vals.ai) |
| [Excel Modeling Benchmark](https://financebenchmark.ai/benchmarks/emb.md) | documentQA, forensic | **31.7%** | 2026-09 | [Paper](https://vals.ai/benchmarks/emb) |
| [Finance Agent Benchmark](https://financebenchmark.ai/benchmarks/finance-agent.md) | agentTasks | **37.7%** | 2026-09 | [Paper](https://vals.ai/benchmarks/fabv2) |
| [Legal Research Bench](https://financebenchmark.ai/benchmarks/legal-research-bench.md) | documentQA, forensic | **15.4%** | 2026-06 | [Paper](https://vals.ai) |
| [MedCode](https://financebenchmark.ai/benchmarks/medcode.md) | verification | **38.6%** | 2026-06 | [Paper](https://vals.ai) |
| [TaxEval v2](https://financebenchmark.ai/benchmarks/taxeval-v2.md) | verification, forensic | **73.1%** | 2026-06 | [Paper](https://vals.ai) |
| [Vibe Code Bench v1.1](https://financebenchmark.ai/benchmarks/vibe-code-bench-v1-1.md) | agentTasks | **7.6%** | 2026-06 | [Paper](https://vals.ai) |
| [Harvey's Legal Agent](https://financebenchmark.ai/benchmarks/harveys-legal-agent.md) | agentTasks, forensic | **0.4%** | 2026-06 | [Paper](https://vals.ai) |
| [GPQA Diamond](https://financebenchmark.ai/benchmarks/gpqa-diamond.md) | numerical, verification | **86.1%** | 2026-06 | [Paper](https://vals.ai) |
| [LiveCodeBench](https://financebenchmark.ai/benchmarks/livecodebench.md) | agentTasks | **86.0%** | 2026-06 | [Paper](https://vals.ai) |
| [LegalBench](https://financebenchmark.ai/benchmarks/legalbench.md) | documentQA, forensic | **82.1%** | 2026-06 | [Paper](https://vals.ai) |
| [MMLU Pro](https://financebenchmark.ai/benchmarks/mmlu-pro.md) | numerical, verification | **85.8%** | 2026-06 | [Paper](https://vals.ai) |
| [ProgramBench](https://financebenchmark.ai/benchmarks/programbench.md) | agentTasks | **0.0%** | 2026-06 | [Paper](https://vals.ai) |
| [SWE-bench](https://financebenchmark.ai/benchmarks/swe-bench.md) | agentTasks | **69.0%** | 2026-06 | [Paper](https://vals.ai) |
| [Terminal-Bench 2.1](https://financebenchmark.ai/benchmarks/terminal-bench-2-1.md) | agentTasks | **50.9%** | 2026-06 | [Paper](https://vals.ai) |

## 4. Certification & Verification {#verification}

Embed live verification badge:

```html
<img src="https://financebenchmark.ai/embed/badge/nemotron-3-ultra" alt="Nemotron 3 Ultra FinanceBenchmark Score" />
```