NVIDIAactiveOpen Source

Nemotron Super 49B

llama-3.3-nemotron-super-49b

NVIDIA-customized Llama 3.3 with dynamic reasoning via Thinking/Non-Thinking modes.

Context Window

131.1K

tokens

Max Output

4.1K

tokens

Input Price

—

per 1M tokens

Output Price

—

per 1M tokens

Details

Familynemotron

Parameters49B

Training Cutoff2024-12-01

ReleasedMarch 18, 2025

FunctionsStreamingJSON ModeCodeReasoningTool Use

HumanEvalFunction-level Python code generation

73.5%

MMLU-ProHarder successor to MMLU

62%

MATH-500Competition-style math

61%

curl pikaainews.com/api/models/nvidia-nemotron-super-49b

npx pika-models info nvidia-nemotron-super-49b

Third-Party Providers & Aggregators

DeepInfra

Lowest per-token rates for open-source models.

Nemotron-70B

nemotron-70b

Nemotron 70B Instruct

llama-3.1-nemotron-70b