llmboard.aiAI model intelligence
Home

Model Rankings

OverallOpen ModelsAgentCodingReasoningMathKnowledgeInstruction FollowingTextVision
Image GenerationImage Editing
Video GenerationImage to VideoVideo Editing
Text to SpeechSpeech to Text
Embeddings

Efficiency

Chat Token PricingImage PricingVideo PricingAudio Pricing
Chat Speed & LatencyProvider Reliability

Benchmarks

GPQAMMLU-ProAIME 2025SWE-Bench VerifiedMMLUHumanity's Last ExamLiveCodeBenchMATHHumanEvalMMMU-Pro
All Benchmarks

Tools

Model DirectoryCompare Models

Scoring & Data

Scoring & Data
393 models668 benchmarks

Leaderboard Center

Overall RankingCodingCore BenchmarksPrice & ValueRuntime Performance

Modalities

All ModelsImage GenerationImage EditingVideo GenerationImage-to-VideoVideo EditingText-to-SpeechSpeech-to-TextEmbeddings

Data & Methods

Scoring MethodAll BenchmarksReasoningMath

Vendors

All VendorsOpenAIAnthropicGoogle
llmboard.aiCopyright 2026 llmboard.ai

Model vendor

NVIDIA

GPU and AI company

Updated Aug 17, 2026. Data as of 2026-08-17

Website
Models38
Versions13
Open defaults15
CountryUS

On this page

  • Models
  • Scores
  • Highlights
  • About
  • FAQ

NVIDIA models

Default-version specifications and scored-version capability values are labeled separately.

30 of 38 rows
Columns

Show columns

Sort by
Model
LLMBoard
Coverage
Context
License
Released
ModelNVNemotron 3 UltraNVIDIALLMBoard62.7Coverage100.0%Context1MLicenseOpen weightsReleasedJun 4, 2026
ModelNVNemotron 3 SuperNVIDIALLMBoard44.3Coverage100.0%Context262.1KLicenseOpen weightsReleasedMar 11, 2026
ModelNVNemotron 3.5 LightningNVIDIALLMBoard37.3Coverage80.0%Context262.1KLicenseOpen weightsReleasedAug 11, 2026
ModelNVNemotron 3 NanoNVIDIALLMBoard34.9Coverage100.0%Context262.1KLicenseOpen weightsReleasedDec 15, 2025
ModelNVNemotron Nano 9BNVIDIALLMBoard31.7Coverage80.0%Context131.1KLicenseOpen weightsReleasedAug 18, 2025
ModelNVLlama 3.1 Nemotron Ultra 253BNVIDIALLMBoard30.0Coverage80.0%ContextN/ALicenseClosed weightsReleasedApr 7, 2025
ModelNVLlama 3.3 Nemotron Super 49BNVIDIALLMBoard22.3Coverage80.0%Context131.1KLicenseOpen weightsReleasedJul 25, 2025
ModelNVLlama 3.1 Nemotron Nano 8BNVIDIALLMBoard9.3Coverage80.0%ContextN/ALicenseClosed weightsReleasedMar 18, 2025
ModelNVLlama 3.1 Nemotron 70BNVIDIALLMBoard2.6Coverage80.0%Context128KLicenseOpen weightsReleasedOct 1, 2024
ModelNVLlama Nemotron Embed VL 1BNVIDIALLMBoardN/ACoverageN/AContext32.8KLicenseOpen weightsReleasedFeb 10, 2026
ModelNVNV-Embed-v2NVIDIALLMBoardN/ACoverageN/AContext32.8KLicenseOpen weightsReleasedSep 9, 2024
ModelNVNV-Embed-v1NVIDIALLMBoardN/ACoverageN/AContext32.8KLicenseOpen weightsReleasedSep 13, 2024
ModelNVnemotron-colembed-vl-8b-v2NVIDIALLMBoardN/ACoverageN/AContext262.1KLicenseOpen weightsReleasedJan 7, 2026
ModelNVllama-nemotron-colembed-vl-3b-v2NVIDIALLMBoardN/ACoverageN/AContext8.2KLicenseOpen weightsReleasedJan 21, 2026
ModelNVnemotron-colembed-vl-4b-v2NVIDIALLMBoardN/ACoverageN/AContext262.1KLicenseOpen weightsReleasedJan 7, 2026
ModelNVllama-nemoretriever-colembed-3b-v1NVIDIALLMBoardN/ACoverageN/AContext8.2KLicenseOpen weightsReleasedJun 27, 2025
ModelNVllama-nemoretriever-colembed-1b-v1NVIDIALLMBoardN/ACoverageN/AContext8.2KLicenseOpen weightsReleasedJun 27, 2025
ModelNVparakeet-tdt-0.6b-v2NVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVcanary-1b-flashNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVcanary-1bNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVparakeet-tdt-1.1bNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVcanary-180m-flashNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVparakeet-rnnt-1.1bNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVcanary-1b-v2NVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVparakeet-tdt_ctc-110mNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVparakeet-rnnt-0.6bNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVparakeet-ctc-0.6bNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVstt_en_conformer_ctc_largeNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVstt_en_fastconformer_ctc_largeNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVstt_en_fastconformer_transducer_largeNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVstt_en_conformer_ctc_smallNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVSana Sprint 1.6BNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVCosmos3-Super-Image2Video-4StepNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVCosmos3-Super-Image2VideoNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVMagpie-Multilingual 357M (Feb 2026)NVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVMagpie-Multilingual 357MNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVMagpie MultilingualNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A
ModelNVCanary Qwen 2.5BNVIDIALLMBoardN/ACoverageN/AContextN/ALicenseClosed weightsReleasedN/A

NVIDIA capability scores

Each unique model uses its latest scored version.

NVIDIA LLMBoard scores

NVIDIA Highlights

Leading NVIDIA models by score, price and context window.

Top scoreNemotron 3 Ultra62.7 LLMBoardLowest official inputNemotron 3 Super$0.20 / 1MLargest contextNemotron 3 Ultra1M tokens
Open-weight defaults15of 38 models

What is NVIDIA?

Background and model catalog information for NVIDIA.

GPU and AI company

Model ownership is separate from API hosting. One model can be offered by multiple API providers with different prices.

Vendor
NVIDIA
Models
38
Versions
13
Country
US
Founded
Not available

Catalog data as of 2026-08-17.

FAQ

Common questions about NVIDIA.

What is NVIDIA's highest-ranked model?

Nemotron 3 Ultra currently has the strongest LLMBoard score from this vendor.

How many NVIDIA models are tracked?

The catalog contains 38 independently selectable NVIDIA models and 13 versions.

Does NVIDIA host every listed model API?

Not necessarily. Vendor identifies the model developer; API providers are tracked separately.

Does NVIDIA publish open-weight models?

15 current default versions are marked as open weight.