Nemotron 3.5 Super VL (120B A12B)
NVIDIAOpen WeightOpenMDW License v1.1 · Commercial OK
Description
NVIDIA Nemotron 3.5 Super VL is an open multimodal 120B total / 12B active Mixture-of-Experts model (hybrid Mamba-2 + Attention with a C-RADIO / RADIO vision encoder) built for long-running agents. It accepts text, image, and video inputs and produces text, with up to 1M context, tool use, and configurable reasoning. Released with open weights under OpenMDW-1.1.
Release Date
2026-10-15
Parameters
120.0B
Context Length
1.0M
Modalities
image, text, video
Capability Radar
64
general
60
coding
50
reasoning
60
scienceest.
40
agents
70
multimodal
Science is estimated from LLM Stats science scores or reasoning when no dedicated science benchmarks are available.
Rankings
| Domain | #Rank | Score | Source |
|---|---|---|---|
| Agentic Capability | 62 | 48.0 | LS |
| Multimodal Ranking | 41 | 59.0 | LS |
Benchmark Scores (LLM Stats)
(LLM Stats (zeroeval))Agents
Tau3 Banking
18.9%SR
General
GDPval-AA
1211.00 / 3000SR
Instruction Following
IFBench
70.9%SR
Math
MathVision
78.2%SR
Multimodal
OSWorld-Verified
51.1%SR
Reasoning
GPQANYU + Cohere + Anthropic (2023)
89.5%SR
LiveCodeBench v6
88.2%SR
CharXiv-R
80.8%SR
BrowseCompOpenAI (2025)
78.0%SR
SWE-Bench Verified
71.1%SR
AA-LCR
68.6%SR
SWE-bench Multilingual
67.7%SR
Terminal-Bench 2.1
66.7%SR
SWE-Bench ProPrinceton NLP (2024)
51.8%SR
SciCode
47.1%SR
Humanity's Last Exam
31.8%SR
Science
OmniScience
26.2%SR
Vision
OCRBench-V2 (en)
68.1%SR
MMLongBench-Doc
0.63 / 100SR
AA Evaluation Indices
(Artificial Analysis)No AA evaluation data available
LLM Stats Category Scores
(LLM Stats (zeroeval))Legal100
Finance100
Agents100
Reasoning87
General64
Search80
Image To Text70
Instruction Following70
Long Context70
Multimodal70
Physics70
Frontend Development70
Biology70
Chemistry70
Code60
Vision60
Math50
Tool Calling40
Knowledge30
Science30
Pricing
Input Price$0 / 1M tokens
Output Price$0 / 1M tokens
Blended Price (3:1)$0 / 1M tokens
Speed
No speed data available
Provider Price Ranking
Provider Price Ranking
2 providers
Cheapest: NVIDIAMost Expensive: DeepInfra
ProviderInputOutput
1NVIDIAPRIMARY
$0
$0
2DeepInfra
$0
$0
Compare pricing across different API providers for this model.