nemotron-3.5-lightning-30b-a3b: LiveBench scores and how to run it with BrainOutput
nemotron-3.5-lightning-30b-a3b is a model from NVIDIA. In the LiveBench release of 2026-06-25 (fetched 2026-10-24) it ranks #71 of 71, with a global score of 44.49.
What is nemotron-3.5-lightning-30b-a3b?
nemotron-3.5-lightning-30b-a3b is listed on LiveBench, a public, continuously refreshed benchmark that scores models on seven categories: reasoning, coding, agentic coding, mathematics, data analysis, language and instruction following. Its strongest category in this snapshot: instruction following (#68 of 71).
We verify the licence and whether its weights are available during the AI diagnostic, before recommending anything.
The model at a glance
| Vendor | NVIDIA |
|---|---|
| Weights | Not publicly stated |
| Licence | We verify the licence during the AI diagnostic |
| Release date (per LiveBench) | — |
| LiveBench identifier | nemotron-3.5-lightning-30b-a3b |
LiveBench scores, dated
Score and rank out of 71 models, per category, in the LiveBench release of 2026-06-25 fetched 2026-10-24.
LiveBench, release 2026-06-25 · livebench.ai · 71 models · fetched 2026-10-24
Overall rank in this LiveBench snapshot.
Global score: the mean of the seven categories.
Gap to #1, Claude Fable 5.1 Max Effort (83.41).
| Category | Score | Rank |
|---|---|---|
| Reasoning | 42.87 | #71 / 71 |
| Coding | 45.76 | #71 / 71 |
| Agentic coding | 19.95 | #70 / 71 |
| Mathematics | 66.55 | #71 / 71 |
| Data analysis | 43.09 | #71 / 71 |
| Language | 38.29 | #71 / 71 |
| Instruction following | 54.94 | #68 / 71 |
| Global (mean of the 7 categories) | 44.49 | #71 / 71 |
Method note: the global score is the mean of the seven category averages, computed by us from LiveBench's published file; it may differ by a few decimals from the site's own display. Equal scores keep LiveBench's file order. Benchmarks measure benchmark tasks — during an AI diagnostic we re-run the comparison on your own workloads.
Running nemotron-3.5-lightning-30b-a3b with BrainOutput
BrainOutput routes each task to a model by confidentiality: sensitive work stays on a local or private model, and everything else goes to a cloud model only when your policy allows it.
Locally, on your hardware
Checked at diagnostic
Possible only if the weights are published: we verify the licence during the AI diagnostic, before any deployment.
In a private cloud
Checked at diagnostic
Possible only if the weights are published — verified during the AI diagnostic.
With your own key (BYOK)
Available
Plug in your own NVIDIA API key: BrainOutput routes non-sensitive work to it when your policy allows; sensitive work never leaves your perimeter.
Frequently asked questions
Is nemotron-3.5-lightning-30b-a3b an open-weight model?
It is not publicly stated: we verify the licence during the AI diagnostic, before recommending anything. The highest-ranked open-weight model in this snapshot is DeepSeek V4.1 Flash Max Effort (#10, 81.11).
Can we use nemotron-3.5-lightning-30b-a3b with BrainOutput?
Yes, through your own NVIDIA API key (BYOK), for non-sensitive work and only when your policy allows it. Sensitive work stays on an open-weight model running locally or in a private cloud.
Where do these scores come from?
From LiveBench (livebench.ai), release 2026-06-25, file fetched 2026-10-24. The global score is the mean of the seven categories, computed by us. The ranking changes monthly; we re-verify at diagnostic time.
Find out whether nemotron-3.5-lightning-30b-a3b fits your work
An AI diagnostic maps your processes, classes what is sensitive, and benchmarks current models against your real tasks — dated numbers, yours this time.
Sources
- LiveBench — leaderboard, release 2026-06-25 — accessed 2026-10-24