Nemotron 3 Models by NVIDIA AI
About
NVIDIA Nemotron 3 is the 2025-2026 open model family covering Nano 30B-A3B, Super 120B-A12B, Content Safety 4B, VoiceChat 12B, and Nano Omni variants for agentic reasoning, safety classification, and multimodal deployment.
Current Variants
Use-when guidance is derived from seed capabilities, context, release, and replacement fields.
Use when the workload needs omni, 262k context, and 30B parameters.
Use when the workload needs moderation, 131k context, and 4B parameters.
Use when the workload needs 12B parameters and multimodal inputs.
Use when the workload needs 1.05m context, 120B parameters, and structured outputs.
Use when the workload needs 256k context, 4.0B parameters, and tool use.
Use when the workload needs structured outputs.
Use when the workload needs 128k context and 49B parameters.
Use when the workload needs safety, 4k context, and 70B parameters.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Nemotron 3 Nano Omni | Use when the workload needs omni, 262k context, and 30B parameters. | 2026-04 | omni262k context30B parameters | Current |
| Nemotron 3 Content Safety | Use when the workload needs moderation, 131k context, and 4B parameters. | 2026-03 | moderation131k context4B parameters | Current |
| Nemotron 3 VoiceChat | Use when the workload needs 12B parameters and multimodal inputs. | 2026-03 | 12B parametersmultimodal inputs | Current |
| Nemotron 3 Super-120B-A12B | Use when the workload needs 1.05m context, 120B parameters, and structured outputs. | 2026-03 | 1.05m context120B parametersstructured outputs | Current |
| Nemotron 3 Nano | Use when the workload needs 256k context, 4.0B parameters, and tool use. | 2025-12 | 256k context4.0B parameterstool use | Current |
| Nemotron 3 Nano 30B-A3B | Use when the workload needs structured outputs. | 2025-12 | structured outputs | Current |
| Llama 3.3 Nemotron Super 49B v1 | Use when the workload needs 128k context and 49B parameters. | 2025-06 | 128k context49B parameters | Current |
| Llama 3.1 Nemotron 70B Reward | Use when the workload needs safety, 4k context, and 70B parameters. | 2024-10 | safety4k context70B parameters | Current |
| Nemotron 3 Ultra | Use when the workload needs 128k context. | 2024-09 | 128k context | Current |
Release Timeline
6 release groupsSpecifications(9 models)
| Model | Released | Context | Parameters | Vision | Multimodal | Fn Calling | Tool Use | Structured Outputs |
|---|---|---|---|---|---|---|---|---|
| Nemotron 3 Nano Omni | 2026-04 | 262k | 30B | No | Yes | No | No | No |
| Nemotron 3 Content Safety | 2026-03 | 131k | 4B | Yes | Yes | No | No | No |
| Nemotron 3 VoiceChat | 2026-03 | — | 12B | Yes | Yes | No | No | No |
| Nemotron 3 Super-120B-A12B | 2026-03 | 1.05m | 120B | No | No | No | No | Yes |
| Nemotron 3 Nano | 2025-12 | 256k | 3.97B | No | No | Yes | Yes | No |
| Nemotron 3 Nano 30B-A3B | 2025-12 | — | 30B (3B active) | No | No | No | No | Yes |
| Llama 3.3 Nemotron Super 49B v1 | 2025-06 | 128k | 49B | No | No | No | No | No |
| Llama 3.1 Nemotron 70B Reward | 2024-10 | 4k | 70B | No | No | No | No | No |
| Nemotron 3 Ultra | 2024-09 | 128k | 550B (55B active) | No | No | No | No | No |
Available From(7 providers)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| Nemotron 3 Nano 30B-A3B | Vercel AI Gateway | $0.05 | $0.24 | Serverless |
| Nemotron 3 Nano 30B-A3B | AWS Bedrock | $0.06 | $0.24 | Serverless |
| Nemotron 3 Super-120B-A12B | OpenRouter | $0.09 | $0.45 | Serverless |
| Nemotron 3 Super-120B-A12B | DeepInfra | $0.1 | $0.5 | Serverless |
| Nemotron 3 Super-120B-A12B | NVIDIA NIM | $0.1 | $0.5 | Serverless |
| Nemotron 3 Super-120B-A12B | Vercel AI Gateway | $0.15 | $0.65 | Serverless |
Frequently Asked Questions
- What is Nemotron 3 used for?
- Nemotron 3 is used for omni, moderation, and safety. The family description and listed model capabilities point to those workloads as the best fit.
- How does Nemotron 3 compare to NVIDIA Nemotron Nano 12B v2 VL?
- Nemotron 3 by NVIDIA AI is strongest where you need omni, while NVIDIA Nemotron Nano 12B v2 VL by NVIDIA AI is the closest related family to check for structured outputs. Nemotron 3 has 9 listed variants and reaches up to 1.05m context, so compare the specs and pricing tables before choosing a production model.
- Which Nemotron 3 model should I use?
- For the lowest listed input price, start with Nemotron 3 Nano 30B-A3B through Vercel AI Gateway at $0.05/1M input tokens. For the most capable/latest local choice, evaluate Nemotron 3 Nano with 256k context and tool use and function calling.
Models(9)
Nemotron 3 Nano Omni
Nemotron 3 Content Safety
Nemotron 3 VoiceChat
Nemotron 3 Super-120B-A12B
Nemotron 3 Nano
Nemotron 3 Nano 30B-A3B
Llama 3.3 Nemotron Super 49B v1
Llama 3.1 Nemotron 70B Reward
Nemotron 3 Ultra






