llmboard.aiLeaderboard Center
Overall
Overall RankingOpen Models
Tools
Model DirectoryCompare Models
Capabilities
CodingReasoningMathKnowledgeInstruction Following
Price & Efficiency
Price & ValueCapability vs. PriceRuntime Performance
Modalities
Image GenerationVideo GenerationSpeech ModelsEmbeddings
Core Benchmarks
GPQAMMLU-ProAIME 2025SWE-Bench VerifiedMMLUHumanity's Last ExamLiveCodeBenchMATHHumanEvalMMMU-ProView all benchmarks
Methods
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

NVIDIA model product

Llama 3.3 Nemotron Super 49B

3-70B-Instruct.

Updated Aug 12, 2026. Default version: Llama-3.3 Nemotron Super 49B v1

Compare
LLMBoard score23.0Llama-3.3 Nemotron Super 49B v1
Coverage80%7 benchmark families
Context window131.1KTokens
Official input priceN/AOfficial price unavailable

On this page

  • Specification
  • Capability
  • Benchmarks
  • Arena
  • Runtime
  • Pricing
  • Versions
  • About
  • Compare
  • Similar models
  • FAQ

Llama 3.3 Nemotron Super 49B Specifications

Technical details for the model's default version.

Version
Llama-3.3 Nemotron Super 49B v1
Released
Mar 18, 2025
Knowledge cutoff
Dec 31, 2023
Parameters
49.9B
Context window
131.1K
Max output
131.1K
Inputs
text
Outputs
text
Open weights
Yes
License
Llama 3.1 Community License

Llama 3.3 Nemotron Super 49B Capability Profile

This profile uses the model's current scored version. Arena ratings and prices are shown separately.

Llama-3.3 Nemotron Super 49B v1 category scores

Llama 3.3 Nemotron Super 49B Benchmark Results

Benchmark scores for Llama-3.3 Nemotron Super 49B v1.

7 rows
Columns

Show columns

MBPP91.3%023396.9%CAug 11, 2026
MT-Bench91.7%021290.9%CAug 11, 2026
BFCL v273.7%03550.0%CAug 11, 2026
Arena Hard88.3%042688.0%CAug 11, 2026
MATH-50096.6%123264.5%CAug 11, 2026
AIME 202558.4%9811414.2%CAug 11, 2026
GPQA66.7%14223439.5%CAug 11, 2026

Llama 3.3 Nemotron Super 49B Arena Results

Preference and agent-evaluation results for the default version.

2 rows
Columns

Show columns

textoverall2141308.52,218N/AAug 10, 2026
text style controloverall2211328.02,218N/AAug 10, 2026

Llama 3.3 Nemotron Super 49B Runtime Performance

Provider-specific output speed and catalog latency for Llama-3.3 Nemotron Super 49B v1. Runtime does not affect the capability score.

No runtime data

No provider-specific speed or latency record is linked to the default version yet.

Browse runtime rankings

Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.

Llama 3.3 Nemotron Super 49B Pricing

Official vendor API pricing appears first, followed by individual provider offers.

Official API
N/A
Official provider
N/A
Lowest third-party
From $0.15 input, $0.15 output per 1M via NanoGPT
Tracked offerings
2
2 rows
Columns

Show columns

Nvidianvidia/llama-3.3-nemotron-super-49b-v1globalN/AN/A131.1KAug 11, 2026
NanoGPTnvidia/Llama-3.3-Nemotron-Super-49B-v1global$0.15$0.15128KAug 11, 2026

Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.

Llama 3.3 Nemotron Super 49B Versions

Available versions of this model. The score column identifies the version used in the overall ranking.

2 rows
Columns

Show columns

Llama 3.3 Nemotron Super 49B v1.5Jul 25, 2025N/AN/A131.1K131.1KYesUnspecified
Llama-3.3 Nemotron Super 49B v1Mar 18, 202523.049.9B131.1K131.1KYesLlama 3.1 Community License

What is Llama 3.3 Nemotron Super 49B?

Key information about Llama 3.3 Nemotron Super 49B and its available data.

3-70B-Instruct. It's post-trained for reasoning, chat, RAG, and tool calling, offering a balance between accuracy and efficiency (optimized for single H100). It underwent multi-phase post-training including SFT and RL (RLOO, RPO).

Data as of 2026-08-11.

Llama 3.3 Nemotron Super 49B vs nearby models

Open a comparison with the three ranked models immediately above and below this model.

Llama 3.3 Nemotron Super 49BvsDevstral MediumLlama 3.3 Nemotron Super 49BvsGPT-4oLlama 3.3 Nemotron Super 49BvsQwen3 VL 4B ThinkingLlama 3.3 Nemotron Super 49BvsQwen2.5 VL 72BLlama 3.3 Nemotron Super 49BvsMiniCPM SALALlama 3.3 Nemotron Super 49BvsLlama 3.3 70B

Models similar to Llama 3.3 Nemotron Super 49B

Recommendations prioritize the same model type and family, then the closest LLMBoard score.

#161+7.5
NV

Llama 3.1 Nemotron Ultra 253B

NVIDIA

30.5 LLMBoard

DetailsCompare
#154+9.4
NV

Nemotron Nano 9B

NVIDIA

32.4 LLMBoard

DetailsCompare
#147+12.2
NV

Nemotron 3 Nano

NVIDIA

35.2 LLMBoard

DetailsCompare
#220-12.6
NV

Llama 3.1 Nemotron Nano 8B

NVIDIA

10.4 LLMBoard

DetailsCompare
#241-19.9
NV

Llama 3.1 Nemotron 70B

NVIDIA

3.1 LLMBoard

DetailsCompare
#109+21.8
NV

Nemotron 3 Super

NVIDIA

44.8 LLMBoard

DetailsCompare

FAQ

Common questions about Llama 3.3 Nemotron Super 49B.

When was Llama 3.3 Nemotron Super 49B released?

Llama 3.3 Nemotron Super 49B's default version was released on Mar 18, 2025.

How much does Llama 3.3 Nemotron Super 49B cost?

No official standard PAYG price is currently available for Llama 3.3 Nemotron Super 49B. The lowest tracked third-party offer starts at $0.15 input and $0.15 output via NanoGPT.

Who created Llama 3.3 Nemotron Super 49B?

Llama 3.3 Nemotron Super 49B was created by NVIDIA.

What is the context window for Llama 3.3 Nemotron Super 49B?

The default version has a 131.1K token context window.

Is Llama 3.3 Nemotron Super 49B open weight?

Yes. The default version is marked as open weight under Llama 3.1 Community License.

How many API providers offer Llama 3.3 Nemotron Super 49B?

2 provider offerings are linked to the default version.

What models should I compare Llama 3.3 Nemotron Super 49B with?

Nearby ranked alternatives include Devstral Medium, GPT-4o, Qwen3 VL 4B Thinking.