impulse

Nemotron 3 Nano 30B A3B

Open weights

NVIDIA Nemotron 3 Nano 30B A3B is a small language MoE model with highest compute efficiency and accuracy for developers to build specialized agentic AI systems.

API id Nemotron-3-Nano-30B-A3BReleased 14 Dec 2025 (Dec 2025)
Input / 1M
$0.05
at DeepInfra
Output / 1M
$0.20
at DeepInfra
Context window
262K
tokens
Max output
262K
tokens per reply
Hosts
2
2 listings
Released
Dec 2025
14 Dec 2025

Pricing

List prices at DeepInfra, in US dollars per million tokens.

TierInput / 1MOutput / 1M
Standard$0.05$0.20
Cached input
50% off input
$0.025

What it costs a month

At DeepInfra list prices, 30 days.

  • Chat assistant5,000 requests a day, 2K in / 0.5K out, cached$27.75
  • RAG over documents2,000 requests a day, 8K in / 0.6K out, cached$25.20
  • Coding agent1,000 requests a day, 40K in / 2K out, cached$51.00
  • Overnight batch20,000 requests a day, 4K in / 0.4K out$168
Use your own numbers

Where to run it

2 hosts serve Nemotron 3 Nano 30B A3B. The comparison is on blended price, 3 parts input to 1 part output.

HostIn / out per 1M
DeepInfraInference
Nemotron-3-Nano-30B-A3B
$0.05 / $0.20
Novita AIInference
nemotron-3-nano-30b-a3b
$0.05 / $0.20

Capabilities

Core

  • Streaming
  • Function calling
  • JSON mode
  • Structured outputs

Multimodal

  • Vision
  • Image input
  • Audio
  • Image generation
  • Video

Advanced

  • Reasoning
  • Tool calling
  • MCP
  • Prompt caching
  • Batch API
  • Embeddings
Takes in
Text
Produces
Text
Features
Prompt caching
Open weights
License: NVIDIA Open Model License
Runs in about 75.8 GB of GPU memory
Runtimes: TGI, vLLM

Cheaper alternatives

Current models that cost less and keep most of what this one can do: tools, vision, reasoning, structured output, caching and enough context.

Same model, cheaper host

No host sells it for less than DeepInfra.

Smaller NVIDIA models

No cheaper NVIDIA model keeps its key capabilities.

Other frontier labs

No cheaper match from another frontier lab.

Open-weight picks

Call it

Using DeepInfra's API.

Python
from litellm import completion

response = completion(
    model="deepinfra/Nemotron-3-Nano-30B-A3B",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Quick answers

How much does Nemotron 3 Nano 30B A3B cost?

On DeepInfra, Nemotron 3 Nano 30B A3B costs $0.05 per million input tokens and $0.20 per million output tokens. Cached input costs $0.025 per million tokens.

Which APIs offer Nemotron 3 Nano 30B A3B?

Nemotron 3 Nano 30B A3B runs on 2 APIs: DeepInfra and Novita AI.

What is Nemotron 3 Nano 30B A3B's context window?

262K tokens, with up to 262K tokens in one reply.

Is Nemotron 3 Nano 30B A3B being retired?

No retirement date is scheduled for it on any API we track.

What are cheaper alternatives to Nemotron 3 Nano 30B A3B?

Models that cost less and keep most of what it can do: Nvidia Nemotron 3.5 Lightning on DeepInfra (3% less). Savings are on a mix of 3 input tokens to 1 output token.

When was Nemotron 3 Nano 30B A3B released?

NVIDIA released it on 14 Dec 2025.

What is Nemotron 3 Nano 30B A3B's API model id?

Use Nemotron-3-Nano-30B-A3B with DeepInfra. Other hosts use their own ids: nemotron-3-nano-30b-a3b on Novita AI.

Sources

Raw catalog entry for Nemotron-3-Nano-30B-A3B on DeepInfra
{
  "max_output_tokens": 262144,
  "max_input_tokens": 262144,
  "input_cost_per_token": 5e-8,
  "output_cost_per_token": 2e-7,
  "cache_read_input_token_cost": 2.5e-8,
  "supports_prompt_caching": true,
  "litellm_provider": "deepinfra",
  "mode": "chat",
  "supports_function_calling": true,
  "supports_tool_choice": true,
  "supports_reasoning": true,
  "supports_vision": false,
  "source": "https://deepinfra.com/pricing"
}