Trinity Mini
33.7B
- input
- $0.00
- output
- $0.00
- Context
- 256Ktokens
- Released
- 6mo ago
- Size
- 30Bparameters
Cost/1M tokens
Open Nemotron omni model combining reasoning with text, vision, and audio
- Multimodal model
- Open weights · 30B
- Accepts images, audio, video
- Reasoning
- Tools
How it compares
Size class
Medium30BAn open-weight model of 25 to 100 billion parameters, for a workstation or a single server GPU.
Open models of a similar size from other labs
Vs. your stack
Add the models you use to see how this one compares.
Sign inAdditional source evidence
- Source model
- nvidia/nemotron-3-nano-omni-30b-a3b-reasoning
- Source family
- nemotron
- Last updated
- 2026-04-28
- Model context
- 256K
- Model output limit
- 66K
- Open weights
- Yes
- File attachments
- Yes
- Reasoning
- Yes
- Tool calling
- Yes
- Temperature control
- Yes
Complete source record (JSON)
{
"id": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning",
"name": "Nemotron 3 Nano Omni 30B A3B Reasoning",
"description": "Open Nemotron omni model combining reasoning with text, vision, and audio",
"family": "nemotron",
"last_updated": "2026-04-28",
"attachment": true,
"reasoning": true,
"tool_call": true,
"temperature": true,
"modalities": {
"input": [
"text",
"image",
"video",
"audio"
],
"output": [
"text"
]
},
"limit": {
"context": 256000,
"output": 65536
},
"release_date": "2026-04-28",
"open_weights": true
}Frequently asked questions
Running locally
About 21.4 GB at Q4, 37.2 GB at Q8 and 69.5 GB at FP16. These are estimates from the 30B parameter count, weights plus about 15% and 0.5 GB for a short context.
Yes on a 24 GB GPU, it fits at Q4 (estimated 21.4 GB). Based on estimates, it does not fit at Q4, Q8 or FP16 on a 16 GB, 12 GB or 8 GB GPU.
Yes. The Q4 estimate is 21.4 GB, and assuming macOS lets the GPU use about 75% of unified memory, the smallest common Mac memory size that fits Q4 is 32 GB (estimate).
Privacy and data
Yes. Nemotron 3 Nano Omni has open weights, so you can self-host it for full control of your data.
Capabilities and specs
The context window is 256K tokens (256,000) and the maximum output is 66K tokens (65,536).
Yes. Input modalities are text, audio, image and video.
Yes, tool calling (function calling) is supported.
Yes, a reasoning mode is supported, but it is not always on.
Quality and benchmarks
The closest alternatives, matched on model kind and release date with no comparable price, are Nex-N2-Pro by Nex AGI, DiffusionGemma 26B A4B IT by Google DeepMind and Gemma 4 12B IT by Google DeepMind.
Nex-N2-ProNemotron 3 Nano Omni vs Nex-N2-ProDiffusionGemma 26B A4B ITNemotron 3 Nano Omni vs DiffusionGemma 26B A4B ITGemma 4 12B ITNemotron 3 Nano Omni vs Gemma 4 12B IT
Versions and lineage
Nemotron 3 Nano Omni has a 256K token (256,000) context versus 128K tokens (128,000) for Nemotron 3 Content Safety, costs $0.00 input and $0.00 output per 1M tokens, and was released April 28, 2026 versus March 16, 2026 for Content Safety.
Nemotron 3 Content SafetyNemotron 3 Nano Omni vs Nemotron 3 Content Safety
Licensing and access
Nemotron 3 Nano Omni has open weights.
The weights are in the Hugging Face repository nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16.
Basics
Nemotron 3 Nano Omni was made by NVIDIA and released on April 28, 2026.
Yes, Nemotron 3 Nano Omni has open weights.
Generated from Vertical Tabs data, updated November 3, 2026.
More in Nemotron
33.7B1BNemotron 3.5 Embed 1B
8BNemotron 3.5 Embed 8B
8BNemotron 3.5 Rerank 8B
Nemotron 3.5 Super VL
120BNemotron 3.5 Super 120B A12B
0.6BNemotron 3.5 ASR Streaming Multilingual 0.6B
30BNemotron 3.5 Lightning
1BNemotron 3 Embed 1B
550BNemotron 3 Ultra
4.3BNemotron 3.5 Content Safety
30B
Discussion
No comments yet
Be the first to share your thoughts.