Inference
EN
NVIDIA

Nemotron 3.5 Lightning 30B A3B NVFP4

Nemotron 3.5 Lightning 30B A3B NVFP4 is a reasoning model from NVIDIA in the Nemotron 3.5 Lightning family. 14 benchmarks count toward its index, in 5 categories.

  • Open weights
  • Reasoning
  • Input: text
  • Released 2026-08
RankedEnough results for an index.
IndexAll categories, out of 100.37.0±6.8Median 51.0
CoverageIndex weight that has results.75%
SpeedOutput tokens per second.—Median 67/s
First tokenSeconds to the first output token.—Median 1.3 s
Input / 1MUS dollars per 1M input tokens.—Median $0.60
Output / 1MUS dollars per 1M output tokens.—Median $2.55
ContextMaximum tokens in one request.1MMedian 262K
EloLMArena rating and rank.—Median 1422

Run locally

31.6B parameters
16-bit
67.4 GB
8-bit
36.3 GB
4-bit
21.1 GB
Weights
Hugging Face (opens in a new tab) ↗

Memory need with a short context. The 4-bit copy fits in 24 GB. Compare local models

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.5 counted30.2201 / 218
CodingCode writing and repair.2 counted38.7182 / 209
ReasoningLogic problems and puzzles.2 counted38.9170 / 227
MultimodalTasks with images and text.None counted——
KnowledgeFacts and expert knowledge.4 counted39.5197 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted57.1—
MathMath problems.None counted——

Results

14 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
Massive Multitask Language Understanding ProfessionalKnowledge81.6%36 / 75
Graduate-Level Google-Proof Q&AKnowledge75.6%78 / 141
PinchBenchAgentic83.4%6 / 10
Artificial Analysis GPQA DiamondKnowledge74.3%122 / 199
Artificial Analysis Humanity's Last ExamKnowledge10.6%143 / 222
Instruction Following BenchmarkInstructionAbove fitFar above this model's index.72.9%37 / 57
GDPval-AA normalizedAgentic6.9%93 / 133
Critical Physics TasksReasoning0.0%150 / 214
Artificial Analysis Long Context ReasoningReasoning49.2%159 / 217
GPQA Diamond · BenchLMKnowledge75.6%68 / 84
Software Engineering Benchmark VerifiedCoding52.8%94 / 114
Scientific Code BenchmarkCoding31.4%30 / 35
Artificial Analysis SciCodeCoding32.1%102 / 114
Artificial Analysis Omniscience AccuracyKnowledge14.7%153 / 164
τ³-Bench Tool-Agent-User EvaluationAgentic9.5%30 / 32
Terminal-Bench 2.1 · BenchLMAgentic23.5%82 / 86
Humanity's Last Exam without toolsKnowledge10.5%51 / 53
BrowseCompAgentic36.8%62 / 63
Artificial Analysis Openness IndexKnowledge83.3%—

Sources

BenchLMCC BY-NC 4.0

History

5 releases
Index37.6 → 37.0
23 Oct 202612 Nov 2026

Questions

How good is Nemotron 3.5 Lightning 30B A3B NVFP4?
Its index is 37.0 out of 100. That is place 228 of 283 ranked models.
What is Nemotron 3.5 Lightning 30B A3B NVFP4 best at?
Knowledge: 39.5 out of 100, place 197 of 268.
What is the context window of Nemotron 3.5 Lightning 30B A3B NVFP4?
1M tokens.
Is Nemotron 3.5 Lightning 30B A3B NVFP4 open source?
Yes. Its weights are published.
When was Nemotron 3.5 Lightning 30B A3B NVFP4 released?
August 2026.

Badge

Nemotron 3.5 Lightning 30B A3B NVFP4 on Inference360: place 228 of 283, index 37.0
HTML
<a href="https://inference360.ai/models/nemotron-3-5-lightning-30b-a3b-nvfp4"><img src="https://inference360.ai/badges/nemotron-3-5-lightning-30b-a3b-nvfp4.svg" alt="Nemotron 3.5 Lightning 30B A3B NVFP4 on Inference360: place 228 of 283, index 37.0" height="20"></a>
Markdown
[![Nemotron 3.5 Lightning 30B A3B NVFP4 on Inference360: place 228 of 283, index 37.0](https://inference360.ai/badges/nemotron-3-5-lightning-30b-a3b-nvfp4.svg)](https://inference360.ai/models/nemotron-3-5-lightning-30b-a3b-nvfp4)

Compare with

Ling 2.6 Flash37.0MiniCPM5-2B37.1DeepSeek V3 032437.1Sarvam 105B36.7Nemotron 3 Ultra50.8

More from NVIDIA

All models
Nemotron 3.5 LightningUnrankedCosmos3-EdgeUnrankedSwitchyardUnrankedNemotron 3.5 Super VL 120B A12B51.0Nemotron 3.5 Super VL 120B A12B NVFP450.8