NVIDIA
Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)
Llama 3.1 Nemotron Ultra 253B v1 (Reasoning) is a Meta Llama 3 family model, part of Meta's widely deployed open-weight LLM series for assistants, retrieval, coding, and self-hosted applications. The profile helps compare this specific size and instruction variant with newer open and proprietary models in the local snapshot.
Introducing Meta Llama 3Operational Metrics
Output Speed41 tok/s
First Token0.71s
Blended Price$0.900/M
Model Metadata
Queryable facts extracted from the upstream model payload.
ReleaseApr 7, 2025
Context Windown/a
Modalitiesn/a
Strength: AIME
Rank #32 across 194 models.
74.7%
Strength: MMLU-Pro
Rank #67 across 345 models.
82.5%
Strength: MATH-500
Rank #41 across 201 models.
95.2%
Watch Area: Speed
Rank #252 across 293 models.
41 tok/s
Watch Area: Coding
Rank #278 across 410 models.
13.1
Strength Profile
Percentile score by analysis domain.
* Cost is inverted: lower input, output, and blended prices rank higher.
Benchmark Percentiles
Higher bars mean stronger relative placement.
All Benchmarks
| Metric | Domain | Value | Rank |
|---|---|---|---|
| Artificial Analysis Intelligence Index | overall | 15.0 | #317 |
| Artificial Analysis Coding Index | coding | 13.1 | #278 |
| Artificial Analysis Math Index | math | 63.7 | #109 |
| MMLU-Pro | reasoning | 82.5% | #67 |