Skip to catalog
FbFable
CatalogLabsScoresCompare
Catalog
N

NVIDIA

Nemotron 3.5 Lightning

nemotron-3.5-lightning

37Index
NVIDIA OpenGenerally available128K context30BText

Pick this when

Local agents and research. The most completely open release in this catalog.

Skip when

You wanted a hosted flagship. This is a 30B you run yourself.

NVIDIA's pitch is reproducibility, not the intelligence board. Believe that pitch.

Price

Input / 1M
$0.05
Output / 1M
$0.20
Measured task
$0.03

Weights, data, recipe, and RL environment are published.

Shape

Context
128K
Max output
16K
Released
2026-08-11
Tools
Yes
Reasoning
Yes

Scores

Intelligence index37
GPQA Diamond52
Terminal-Bench 2.127
Agentic index17
Arena
—
Speed
95/s
GPQA
52.4%
Term.
26.8%
  • AA Index. The single number most people mean by “how smart.” It is a blend, so a specialist can lose here and still win the job you care about.
  • GPQA. A clean test of hard reasoning. PhD experts sit around 65%. It says little about writing, tools, or taste.
  • Term.. Closest public proxy for coding agents that live in a terminal. A high index score with a weak terminal score is a warning.
  • Arena. Captures taste and usefulness that unit tests miss. Sample size varies by model, so treat gaps under ~100 Elo as noise.

NVIDIA · Santa Clara. Nemotron publishes weights, data, and the training recipe. 3.5 Lightning is a 30B local agent, not a board leader. It is the most open release here.

Compiled 2026-08-28 from public Artificial Analysis, LMSYS-style arena, and first-party price sheets. Scores move. Recheck before you spend.

A two-point gap on the intelligence index is noise for most work. Cost per finished task is the number that shows up on the invoice.