NVIDIA Nemotron 3 Nano 30B A3B (Reasoning) is a leading model in intelligence. It is notably fast, though it tends to be very verbose. The model supports text input, generates text output, and features a 1 million token context window.
Author
NVIDIARelease Date
2025-12-14
Knowledge Cutoff
2025-11-28
License
Proprietary
I/O Format
Context Length
262K / 228K
API I/O (1M)
$0.05 / $0.2
How to Use
—
Output Speed
134 tok/sArena Overall
1317Intelligence Index
24.3Coding Index
19.0Math Index
91.0LiveBench
—ForecastBench
—GPQA Diamond
75.7%HLE
10.2%MMLU-Pro
79.4%AIME 2025
91.0%MATH-500
—LB Reasoning
—LB Math
—LB Data Analysis
—LiveCodeBench
74.1%LB Coding
—LB Agentic
—TAU2
40.9%TerminalBench
13.6%SciCode
29.6%IFBench
71.1%AA-LCR
0.3Hallucination (HHEM)
—Factual Consistency (HHEM)
—LB Language
—LB Instruction Following
—1 / 3
Swipe to compare
NVIDIA Nemotron 3 Nano 30B A3B (Reasoning) is a leading model in intelligence. It is notably fast, though it tends to be very verbose. The model supports text input, generates text output, and features a 1 million token context window.
Author
NVIDIARelease Date
2025-12-14
Knowledge Cutoff
2025-11-28
License
Proprietary
I/O Format
Context Length
262K / 228K
API I/O (1M)
$0.05 / $0.2
How to Use
—
Output Speed
134 tok/sArena Overall
1317Intelligence Index
24.3Coding Index
19.0Math Index
91.0LiveBench
—ForecastBench
—GPQA Diamond
75.7%HLE
10.2%MMLU-Pro
79.4%AIME 2025
91.0%MATH-500
—LB Reasoning
—LB Math
—LB Data Analysis
—LiveCodeBench
74.1%LB Coding
—LB Agentic
—TAU2
40.9%TerminalBench
13.6%SciCode
29.6%IFBench
71.1%AA-LCR
0.3Hallucination (HHEM)
—Factual Consistency (HHEM)
—LB Language
—LB Instruction Following
—