# Nemotron 3 Nano 4B

Nemotron 3 Nano 4B is an AI model released by NVIDIA on Mar 16 2026. It has 4B parameters, a 256k context window and open weight. Tracked results include 53.2% on GPQA Diamond, 39.8% on τ²-bench (retail) and 33.3% on τ²-bench (airline).

## Facts

| Field | Value |
| --- | --- |
| Model | Nemotron 3 Nano 4B |
| Developer | NVIDIA |
| Release date | Monday, Mar 16 2026 |
| Licensing | Open Weight |
| Parameters | 4B |
| Context window | 256k |

## Tracked benchmark scores

| Benchmark | Score | Source | What it measures |
| --- | --- | --- | --- |
| τ²-bench (retail) | 39.8% | Lab | task success in the retail domain of the second-generation customer-service evaluation |
| τ²-bench (airline) | 33.3% | Lab | task success in the airline domain of the second-generation customer-service evaluation |
| τ²-bench (telecom) | 33% | Lab | task success in the telecom domain of the second-generation customer-service evaluation |
| MATH-500 | 95.4% | Lab | accuracy on the 500-problem MATH subset |
| AIME 2025 | 78.5% | Lab | single-answer accuracy on the 2025 invitational mathematics exam without tools |
| GPQA Diamond | 53.2% | Lab | Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. |

## About Nemotron 3 Nano 4B

Nemotron 3 Nano 4B was released in March 2026 for local conversational agents. NVIDIA compressed Nemotron Nano 2’s 9B model and trained the smaller hybrid checkpoint for instruction following and tool use.

NVIDIA distributed BF16, FP8 and GGUF weights for deployment on devices including RTX PCs and Jetson. The benchmark scores shown here use BF16 with reasoning enabled, including 53.2% on GPQA Diamond and 95.4% on MATH-500.

## Questions and answers

### When was Nemotron 3 Nano 4B released?

Nemotron 3 Nano 4B was released by NVIDIA on Monday, Mar 16 2026.

### Who made Nemotron 3 Nano 4B?

Nemotron 3 Nano 4B was built by NVIDIA. The dominant AI chipmaker, publishing the open Nemotron model family. Ships weights, datasets and training recipes to seed demand for its hardware rather than to sell a chatbot.

### What benchmark scores did Nemotron 3 Nano 4B get?

Nemotron 3 Nano 4B reports 6 tracked benchmark scores — τ²-bench (retail): 39.8%; τ²-bench (airline): 33.3%; τ²-bench (telecom): 33%; MATH-500: 95.4%; AIME 2025: 78.5%; GPQA Diamond: 53.2%. Tracked scores may come from lab reports or independent benchmarks; source details accompany the benchmark data.

### What is the context window of Nemotron 3 Nano 4B?

Nemotron 3 Nano 4B has a context window of 256k. Providers may specify separate input and output token limits.

### How many parameters does Nemotron 3 Nano 4B have?

Nemotron 3 Nano 4B is reported at 4B parameters.

### Is Nemotron 3 Nano 4B open source?

Partly. Nemotron 3 Nano 4B is an open-weight model: the trained weights are free to download and run locally or on your own infrastructure, but the training data and code are not fully released and the license may restrict some commercial uses. It is not open source in the strict sense.

### What came before and after Nemotron 3 Nano 4B?

NVIDIA's previous tracked release was Nemotron 3 Super on Mar 11 2026, 5 days earlier. It was followed by Nemotron 3 Ultra on Jun 4 2026.


---

Canonical page: https://aireleasetracker.com/model/nvidia/nemotron-3-nano-4b
Site index: https://aireleasetracker.com/llms.txt
Source: AI Release Tracker (https://aireleasetracker.com). Most benchmark scores come from lab launch material; gathered results identify the leaderboard that published them.
