‹ PublicAI Index
The LLM benchmark aggregator.
Nemotron 3 Nano 4B
NVIDIA · 4B
Strongest in Knowledge work (#173 of 198, on 1 of its 2 boards), weakest in Agents (#245 of 289). Above par in 0 of 12 scopes. Among the models it meets almost everywhere, it finishes behind GPT-5.6 Terra and Claude Sonnet 5 and ahead of Gemma 4 31B and Mistral Large 3.
§ 1 · Profile
What it is good at
Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.
Core abilities41.2−24#205/2251/4
General intelligence37.3−31.5#200/2141/3
Agents42.8−23.1#245/2891/6
Knowledge work35.6−36.2#173/1981/2
Tool use46.2−27.8—/81✱0/1
Reasoning47.4−17.4—/189✱0/5
Mathematics47.4−16.6—/149✱0/2
Science41.9−21.6—/122✱0/1
Expert reasoning41.6−19.3—/0✱✱
Coding45.4−21—/176✱0/6
Agentic coding45.2−21.8—/169✱0/5
Code generation32.4−32.9—/20✱0/1
§ 2 · Head to head
What it beats, and what beats it
The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for Nemotron 3 Nano 4B, left for the other.
§ 3 · Sources
Where the numbers come from
3 publications, 8 figures. Every one links to the page it was read from.
Artificial Analysis Intelligence Index 5
GDPval-AA -58
Terminal-Bench 2.1 3.7%SciCode 16.4%GPQA Diamond 51.3%HMMT Feb 2026 34.7%HLE 4.9%BFCL v4 36.8%
Badge
[](https://publicai.io/model-index/m/nemotron-3-nano-4b)