Inference
EN
NVIDIA

Nemotron 3.5 Super VL 120B A12B NVFP4

Nemotron 3.5 Super VL 120B A12B NVFP4 is a reasoning model from NVIDIA in the Nemotron 3.5 Super VL family. 18 benchmarks count toward its index, in 6 categories.

  • Open weights
  • Reasoning
  • Input: text
  • Released 2026-10
RankedEnough results for an index.
IndexAll categories, out of 100.50.8±5.3Median 51.0
CoverageIndex weight that has results.85%
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

124.5B parameters
16-bit
262.4 GB
8-bit
139.9 GB
4-bit
80.3 GB
Weights
Hugging Face (opens in a new tab) ↗

Memory need with a short context. The 4-bit copy fits in 128 GB. Compare local models

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.4 counted45.1161 / 218
CodingCode writing and repair.4 counted55.196 / 209
ReasoningLogic problems and puzzles.2 counted55.682 / 227
MultimodalTasks with images and text.4 counted47.171 / 84
KnowledgeFacts and expert knowledge.3 counted52.7116 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.1 counted55.8—
MathMath problems.None counted——

Results

18 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
LiveCodeBench v6Coding87.4%16 / 59
Critical Physics TasksReasoning10.6%63 / 214
Artificial Analysis Humanity's Last ExamKnowledge30.7%74 / 222
GPQA Diamond · BenchLMKnowledge89.9%31 / 84
GDPval-AA normalizedAgentic33.1%55 / 133
BrowseCompAgentic77.5%32 / 63
Artificial Analysis SciCodeCoding47.8%61 / 114
Software Engineering Benchmark VerifiedCoding71.1%64 / 114
Artificial Analysis Long Context ReasoningReasoning67.7%127 / 217
Artificial Analysis Omniscience AccuracyKnowledge25.9%100 / 164
Humanity's Last ExamKnowledge29.0%50 / 77
Instruction Following BenchmarkInstruction71.5%39 / 57
CharXiv ReasoningMultimodal79.9%35 / 49
Humanity's Last Exam without toolsKnowledge31.4%38 / 53
Terminal-Bench 2.1 · BenchLMAgentic60.8%64 / 86
V*Multimodal86.0%11 / 14
SWE-bench ProCoding51.3%67 / 83
τ³-Bench Tool-Agent-User EvaluationAgentic17.8%27 / 32
MathVisionMultimodal77.8%22 / 24
BabyVisionMultimodal29.5%11 / 11
Apex ShortlistMath89.4%—
τ²-Bench Airline DomainAgentic75.8%—
MMLongBench-DocMultimodal62.8%—

Sources

BenchLMCC BY-NC 4.0

History

5 releases
Index51.8 → 50.8
23 Oct 202612 Nov 2026

Questions

How good is Nemotron 3.5 Super VL 120B A12B NVFP4?
Its index is 50.8 out of 100. That is place 145 of 283 ranked models.
What is Nemotron 3.5 Super VL 120B A12B NVFP4 best at?
Reasoning: 55.6 out of 100, place 82 of 227.
What is the context window of Nemotron 3.5 Super VL 120B A12B NVFP4?
1M tokens.
Is Nemotron 3.5 Super VL 120B A12B NVFP4 open source?
Yes. Its weights are published.
When was Nemotron 3.5 Super VL 120B A12B NVFP4 released?
October 2026.

Badge

Nemotron 3.5 Super VL 120B A12B NVFP4 on Inference360: place 145 of 283, index 50.8
HTML
<a href="https://inference360.ai/models/nemotron-3-5-super-vl-120b-a12b-nvfp4"><img src="https://inference360.ai/badges/nemotron-3-5-super-vl-120b-a12b-nvfp4.svg" alt="Nemotron 3.5 Super VL 120B A12B NVFP4 on Inference360: place 145 of 283, index 50.8" height="20"></a>
Markdown
[![Nemotron 3.5 Super VL 120B A12B NVFP4 on Inference360: place 145 of 283, index 50.8](https://inference360.ai/badges/nemotron-3-5-super-vl-120b-a12b-nvfp4.svg)](https://inference360.ai/models/nemotron-3-5-super-vl-120b-a12b-nvfp4)

Compare with

Nemotron 3 Ultra50.8Ling 3.0 Flash50.9GPT-5.4 nano50.9Nemotron 3.5 Super VL 120B A12B51.0Qwen3.6-35B-A3B50.6

More from NVIDIA

All models
Nemotron 3.5 Super VL 120B A12B51.0Nemotron 3.5 Bolt 74B-A9B46.4Nemotron 3.5 Bolt 74B-A9B BF16UnrankedNemotron 3.5 Super VLUnrankedNemotron 3.5 Bolt E4BUnranked