The AI arena is free today

Open Superagent

Llama 3.1 Nemotron Ultra 253B v1 vs Llama 3.3 70B Instruct

Llama 3.1 Nemotron Ultra 253B v1 and Llama 3.3 70B Instruct are closely matched at 19.1 and 13.8 on the LLM Stats Score.

NVIDIA · Meta · Updated for 2026

Which is better?

Llama 3.1 Nemotron Ultra 253B v1 and Llama 3.3 70B Instruct are closely matched on the overall LLM Stats Score at 19.1 and 13.8.

In the 3 individual benchmarks reported for both models, Llama 3.3 70B Instruct wins 2; this is a narrower head-to-head signal than the composite indexes.

Based on current LLM Stats indexes, shared benchmarks, pricing, and model metadata for 2026.

Choose Llama 3.1 Nemotron Ultra 253B v1

  • your work emphasizes reasoning — it leads those capability indexes
  • you want the most recent training data — it shipped Apr 2025

Choose Llama 3.3 70B Instruct

  • you value its reported benchmark strengths — it wins 2 of 3 exact shared results

At a glance

The differences that matter most.

Llama 3.1 Nemotron Ultra 253B v1
NVIDIA
Llama 3.3 70B Instruct
Meta
Core performance indexes
LLM Stats Score
19.1
#209
13.8
#245
Reasoning
18.8
#208
11.7
#254
Coding
12.4
#149
9.3
#176
Cost, coverage & limits
Benchmark wins
1 of 3
2 of 3
Input price
— / M
$0.10 / M
Output price
— / M
$0.20 / M
Context window
—
131,072

Capability indexes

Additional strengths measured across groups of related public benchmarks

1 shared
Index
Llama 3.1 Nemotron Ultra 253B v1
Llama 3.3 70B Instruct
Math
15.6#215
17.5#194
Conservative TrueSkill rating · higher is betterHow scores work

Individual benchmarks

6 reported for Llama 3.1 Nemotron Ultra 253B v1 · 9 for Llama 3.3 70B Instruct

3 shared

Llama 3.1 Nemotron Ultra 253B v1 outperforms in 1 benchmarks (GPQA), while Llama 3.3 70B Instruct is better at 2 benchmarks (BFCL v2, IFEval).

Llama 3.3 70B Instruct shows notably better performance in the majority of benchmarks.

Tue Sep 22 2026 • llm-stats.com

Human preference

Blind head-to-head votes and playground preference scores

Model Size

Parameter count comparison

183.0B diff

Llama 3.1 Nemotron Ultra 253B v1 has 183.0B more parameters than Llama 3.3 70B Instruct, making it 261.4% larger.

NVIDIA
Llama 3.1 Nemotron Ultra 253B v1
253.0Bparameters
Meta
Llama 3.3 70B Instruct
70.0Bparameters
253.0B
Llama 3.1 Nemotron Ultra 253B v1
70.0B
Llama 3.3 70B Instruct

Context Window

Maximum input and output token capacity

Only Llama 3.3 70B Instruct specifies input context (131,072 tokens). Only Llama 3.3 70B Instruct specifies output context (131,072 tokens).

NVIDIA
Llama 3.1 Nemotron Ultra 253B v1
Input- tokens
Output- tokens
Meta
Llama 3.3 70B Instruct
Input131,072 tokens
Output131,072 tokens
Tue Sep 22 2026 • llm-stats.com

License

Usage and distribution terms

Llama 3.1 Nemotron Ultra 253B v1 is licensed under Llama 3.1 Community License, while Llama 3.3 70B Instruct uses Llama 3.3 Community License Agreement.

License differences may affect how you can use these models in commercial or open-source projects.

Llama 3.1 Nemotron Ultra 253B v1

Llama 3.1 Community License

Open weights

Llama 3.3 70B Instruct

Llama 3.3 Community License Agreement

Open weights

Release Timeline

When each model was launched

Llama 3.1 Nemotron Ultra 253B v1 was released on 2025-04-07, while Llama 3.3 70B Instruct was released on 2024-12-06.

Llama 3.1 Nemotron Ultra 253B v1 is 4 months newer than Llama 3.3 70B Instruct.

Llama 3.1 Nemotron Ultra 253B v1

Apr 7, 2025

1.5 years ago

4mo newer
Llama 3.3 70B Instruct

Dec 6, 2024

1.8 years ago

Knowledge Cutoff

When training data ends

Llama 3.1 Nemotron Ultra 253B v1 has a documented knowledge cutoff of 2023-12-01, while Llama 3.3 70B Instruct's cutoff date is not specified.

We can confirm Llama 3.1 Nemotron Ultra 253B v1's training data extends to 2023-12-01, but cannot make a direct comparison without Llama 3.3 70B Instruct's cutoff date.

Llama 3.1 Nemotron Ultra 253B v1

Dec 2023

Llama 3.3 70B Instruct

—

Outputs Comparison

Notice missing or incorrect data?Start an Issue discussion→

Judge for yourself.

Run your own prompts against Llama 3.1 Nemotron Ultra 253B v1 and Llama 3.3 70B Instruct side-by-side, then vote on the output you prefer.

Llama 3.1 Nemotron Ultra 253B v1
✓ Preferred
Llama 3.3 70B Instruct
Open in Playground

FAQ

Common questions about Llama 3.1 Nemotron Ultra 253B v1 vs Llama 3.3 70B Instruct.

Which is better, Llama 3.1 Nemotron Ultra 253B v1 or Llama 3.3 70B Instruct?

Llama 3.1 Nemotron Ultra 253B v1 and Llama 3.3 70B Instruct are closely matched on the LLM Stats Score at 19.1 and 13.8. Llama 3.1 Nemotron Ultra 253B v1 is made by NVIDIA and Llama 3.3 70B Instruct is made by Meta. The best choice depends on your use case — compare their capability indexes, individual benchmarks, pricing, and limits above.

How does Llama 3.1 Nemotron Ultra 253B v1 compare to Llama 3.3 70B Instruct in benchmarks?

Llama 3.1 Nemotron Ultra 253B v1 scores MATH-500: 97.0%, IFEval: 89.5%, GPQA: 76.0%, BFCL v2: 74.1%, AIME 2025: 72.5%. Llama 3.3 70B Instruct scores IFEval: 92.1%, MGSM: 91.1%, HumanEval: 88.4%, MBPP EvalPlus: 87.6%, MMLU: 86.0%.

What are the context window sizes for Llama 3.1 Nemotron Ultra 253B v1 and Llama 3.3 70B Instruct?

Llama 3.1 Nemotron Ultra 253B v1 supports an unknown number of tokens and Llama 3.3 70B Instruct supports 131K tokens. A larger context window lets you process longer documents, conversations, or codebases in a single request.

What are the main differences between Llama 3.1 Nemotron Ultra 253B v1 and Llama 3.3 70B Instruct?

Key differences include LLM Stats Score (19.1 vs 13.8), licensing (Llama 3.1 Community License vs Llama 3.3 Community License Agreement). See the full comparison above for benchmark-by-benchmark results.

Who makes Llama 3.1 Nemotron Ultra 253B v1 and Llama 3.3 70B Instruct?

Llama 3.1 Nemotron Ultra 253B v1 is developed by NVIDIA and Llama 3.3 70B Instruct is developed by Meta.

Related comparisons