Nemotron Cascade 2 30B A3B is an open-source chat model created by NVIDIA. It is designed to process text inputs and generate text outputs, with built-in capabilities for web search, code execution, function calling, extended reasoning, and structured output. The model is intended for tasks requiring agentic behavior and complex problem-solving, operating in both thinking and standard instruction modes.
Input
Output
Context
1M
Max Output
33K
Parameters
30B
Input Modalities
Output Modalities
Some values come from web research and may be incorrect. Report incorrect data
None the catalog can show from stated facts or measured scores.
Bands compare this model with every model the same suite measured in our catalog.
Public resources
Specification sources
As of Nov 10, 2026.
Answered from the facts on this page, with their sources.
Nemotron Cascade 2 30B A3B has a context window of 1,000,000 tokens and a maximum output of 32,768 tokens (source: Web research). A provider may serve less; Inferbase routes on what each provider serves.
Yes. Nemotron Cascade 2 30B A3B is an open-weight model under the Other license. You can download and self-host it within the license terms.
About 30B, read from the model's name. NVIDIA has not published an exact count.
Limitations: Below average on long-context reasoning among measured models. Each is computed from the model's stated facts and benchmark results.
Nemotron Cascade 2 30B A3B accepts Text and produces Text. The Capabilities card lists the API features it supports, such as function calling and structured output.
Nemotron Cascade 2 30B A3B has published results from Artificial Analysis. The Benchmarks card shows each score as reported, with its band among the models the same source measured. Scores are never combined across sources.
Not on Inferbase Managed today. Add your own provider key or endpoint that serves Nemotron Cascade 2 30B A3B, and Inferbase runs it on your account with the same routing and logs.
NVIDIA has not published a parameter count for Nemotron Cascade 2 30B A3B, so the catalog cannot compute its memory. The GPU planner can size it from a count you enter.
Start on the free plan with 5,000 requests a month, no card required.