Language model
Nemotron Nano 12B v2 VL
NVIDIA
- Released
- October 28, 2025
- Data date
- September 26, 2026
Nemotron Nano 12B v2 VL is the NVIDIA BF16 checkpoint for multi-image reasoning, video understanding, and document intelligence. Its vision encoder connects image and video features to the Nemotron Nano language model. Visual question answering, OCR-oriented document work, and summarization are core uses rather than add-on capabilities.
The 12.6-billion-parameter checkpoint processes text, up to four images, and videos through individual frames. NVIDIA lists English as the supported language and a combined 128K-token input and output limit. Video frames must be extracted first because the published inference example does not include a video decoder.
Specifications and access
| Specification | Value and source |
|---|---|
| Weights | Open weightsSource |
| Access | APISource |
| License | NVIDIA Open Model LicenseSource |
| Reasoning | YesSource |
| Tool calling | YesSource |
| Parameters | 12.6 billionSource |
| Combined input and output limit | 128K tokensSource |
| Packaged text configuration | 131,072 tokensSource |
| Inputs | Image, video, textSource |
| Outputs | TextSource |
Popularity
15.1 / 100
Sources and data date
| Type | Evidence and data date |
|---|---|
| Additional source | models.dev catalog (retrieved September 26, 2026) |
| Additional source | NVIDIA Nemotron Nano 12B v2 VL BF16 model card (retrieved October 2, 2026) · NVIDIA Nemotron Nano 12B v2 VL BF16 model card · Editorial description reviewed October 2, 2026 |
| Additional source | NVIDIA Nemotron Nano 12B v2 VL BF16 packaged text configuration (retrieved October 2, 2026) |
| Additional source | NVIDIA Nemotron vision model announcement · Editorial description reviewed October 2, 2026 |
| Additional source | OpenRouter: Popularity index (derived), CC BY 4.0 |
| Additional source | Vercel AI Gateway: Popularity index (derived), CC BY 4.0 |
| Additional source | Creative Commons Attribution 4.0 |