Llama 3.1 Nemotron Nano 4B v1.1 vs NVIDIA Nemotron 3 Nano 4B

Wins 1 of 3 benchmarks
Wins 0 of 3 benchmarks
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models
Llama 3.1 Nemotron Nano 4B v1.15.2
NVIDIA Nemotron 3 Nano 4B4.9
Terminal-Bench 4.0 Terminal-Bench 4.0 · source setting 2: Agents use a terminal to complete executable coding, system administration and research tasks in the 4.0 suite
Llama 3.1 Nemotron Nano 4B v1.1—
NVIDIA Nemotron 3 Nano 4B0
AutomationBench-AA AutomationBench-AA: Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation
Llama 3.1 Nemotron Nano 4B v1.1—
NVIDIA Nemotron 3 Nano 4B0.4

Specs

SpecLlama 3.1 Nemotron Nano 4B v1.1NVIDIA Nemotron 3 Nano 4B
Context128K262K
WeightsOpenOpen
ReleasedMay 2025Mar 2026
Memory (4-bit)2.3 GB2 GB
Licencenvidia-open-model-licensenvidia-nemotron-open-model-license
Commercial useRestrictedRestricted
See images——
Read PDFs??
Hear audio——
Watch video——
Call tools✓✓
Force a tool call??
Return strict JSONPartialPartial
Search the web??
Run code??
Use a computer??

What breaks when you switch

Llama 3.1 Nemotron Nano 4B v1.1 → NVIDIA Nemotron 3 Nano 4B
Nothing known.
NVIDIA Nemotron 3 Nano 4B → Llama 3.1 Nemotron Nano 4B v1.1
Context: 262K → 128K
Memory (4-bit): 2 GB → 2.3 GB