NVIDIALanguage model
Llama 3.1 Nemotron 70B Instruct
Llama 3.1 Nemotron 70B Instruct has no overall index value yet because there are too few independent benchmark results.
PopularityRank 423 of 785
35out of 100-1 vs. last month
based on search, usage and attention
Summary
Llama 3.1 Nemotron 70B Instruct is NVIDIA’s post-trained Llama 3.1 variant with 70 billion parameters and a 128,000-token context.
The model handles text and supports tool calls. Its open release uses the Llama 3.1 Community License rather than NVIDIA's Open Model License. A product deployment therefore follows different usage terms from NVIDIA's Open Model License. This is an instruct checkpoint, neither a base model nor a later Nemotron generation.
All benchmark results
Every published result with its source, test setup and comparison models.
Show all 46 published benchmark results
Page 1 of 4
Measurements without matching peer values
These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.
CritPt v2: Overall (Show measurement, test conditions, and source)
- Source value
- 0
- Score
- 0
- Metric
- pass@1 accuracy
- Unit
- %
- Benchmark version
- v2
- Category
- reasoning
- Task
- overall
- Task label
- Overall
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- Artificial Analysis
- Status
- disputed
- Retrieved at
- 2026-11-02
- Methodology
- Exact AA configuration: Llama 3.1 Nemotron Instruct 70B (release llama-3-1-nemotron-instruct-70b, row llama-3-1-nemotron-instruct-70b, UUID 7393c56a-ec31-48e9-b804-c04f2d2cb641, reasoning=false, effort=not specified, suffix=not specified). AA marks this model variant as non-reasoning and reports 0 reasoning tokens. The full CritPt v2 run covers 70 questions with five repeats. The official grading server reports pass@1.. Source model ID: artificial-analysis/7393c56a-ec31-48e9-b804-c04f2d2cb641/llama-3-1-nemotron-instruct-70b/llama-3-1-nemotron-instruct-70b. Source model name: Llama 3.1 Nemotron 70B Instruct. Provider reported by source: NVIDIA. Model snapshot: Llama 3.1 Nemotron Instruct 70B. Evaluator: Artificial Analysis. Dataset type: artificial-analysis-evaluation. Tool policy: Python functions, symbolic expressions, and numerical answers; official grading server. Attempts: 70 questions, five repeats per question, pass@1. Reasoning: non-reasoning. Reasoning effort: none. Score display: 0.0%. Source ID: artificial-analysis. Evidence capture ID: aa-model-page-2026-11-02-67d46be20a5b. Source payload hash: 67d46be20a5b412db6cfb60ad190441fcaa5513500b8fd47d033168d0e97b6b2. Parser version: artificial-analysis-rsc-models-v3. Imported at: 2026-11-02. Source cohort: artificial-analysis-critpt-v2:v2:standalone:non-reasoning:effort=unspecified:suffix=unspecified
- Comparison methodology
- Exact AA configuration: Llama 3.1 Nemotron Instruct 70B (release llama-3-1-nemotron-instruct-70b, row llama-3-1-nemotron-instruct-70b, UUID 7393c56a-ec31-48e9-b804-c04f2d2cb641, reasoning=false, effort=not specified, suffix=not specified). AA marks this model variant as non-reasoning and reports 0 reasoning tokens. The full CritPt v2 run covers 70 questions with five repeats. The official grading server reports pass@1.. Tool policy: Python functions, symbolic expressions, and numerical answers; official grading server. Attempts: 70 questions, five repeats per question, pass@1. Reasoning: non-reasoning. Reasoning effort: none
- Context
- This AA source field contains the original CritPt score. The earlier CritPt v2 attribution was incorrect; the value and former label remain as disputed historical evidence.. The source dataset marks this result as disputed. It remains visible as documented evidence, but is not reliable comparative proof.
IFBench: Overall (Show measurement, test conditions, and source)
- Source value
- 30.748299319727902
- Score
- 30.748299319727902
- Metric
- IFBench prompt-level loose accuracy (pass@1)
- Unit
- %
- Benchmark version
- AA test set: 294 questions
- Category
- general
- Task
- overall
- Task label
- Overall
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- Artificial Analysis
- Status
- superseded
- Retrieved at
- 2026-11-02
- Methodology
- Exact AA configuration: Llama 3.1 Nemotron Instruct 70B (release llama-3-1-nemotron-instruct-70b, row llama-3-1-nemotron-instruct-70b, UUID 7393c56a-ec31-48e9-b804-c04f2d2cb641, reasoning=false, effort=not specified, suffix=not specified). AA marks this model variant as non-reasoning and reports 0 reasoning tokens. Artificial Analysis runs 294 questions from the IFBench test set with five repeats per question. The official evaluator scores prompt-level accuracy in loose mode as pass@1.. Source model ID: artificial-analysis/7393c56a-ec31-48e9-b804-c04f2d2cb641/llama-3-1-nemotron-instruct-70b/llama-3-1-nemotron-instruct-70b. Source model name: Llama 3.1 Nemotron 70B Instruct. Provider reported by source: NVIDIA. Model snapshot: Llama 3.1 Nemotron Instruct 70B. Evaluator: Artificial Analysis. Dataset type: artificial-analysis-evaluation. Tool policy: Single-turn answer evaluated with the official IFBench source code. Attempts: pass@1, 5 repeats per question. Reasoning: non-reasoning. Reasoning effort: none. Score display: 30.7%. Source ID: artificial-analysis. Evidence capture ID: aa-model-page-2026-11-02-67d46be20a5b. Source payload hash: 67d46be20a5b412db6cfb60ad190441fcaa5513500b8fd47d033168d0e97b6b2. Parser version: artificial-analysis-rsc-models-v3. Imported at: 2026-11-02. Source cohort: artificial-analysis-ifbench:AA test set: 294 questions:standalone:non-reasoning:effort=unspecified:suffix=unspecified
- Comparison methodology
- Exact AA configuration: Llama 3.1 Nemotron Instruct 70B (release llama-3-1-nemotron-instruct-70b, row llama-3-1-nemotron-instruct-70b, UUID 7393c56a-ec31-48e9-b804-c04f2d2cb641, reasoning=false, effort=not specified, suffix=not specified). AA marks this model variant as non-reasoning and reports 0 reasoning tokens. Artificial Analysis runs 294 questions from the IFBench test set with five repeats per question. The official evaluator scores prompt-level accuracy in loose mode as pass@1.. Tool policy: Single-turn answer evaluated with the official IFBench source code. Attempts: pass@1, 5 repeats per question. Reasoning: non-reasoning. Reasoning effort: none
- Context
- The often cited 58 refers to verifiable out-of-domain constraint types, not the number of questions in AA's test run.. This result is retained as historical evidence and is not interchangeable with current results.
IFBench: Overall (Show measurement, test conditions, and source)
- Source value
- 30.748299319727902
- Score
- 30.748299319727902
- Metric
- IFBench prompt-level loose accuracy (pass@1)
- Unit
- %
- Benchmark version
- AA test set: 294 questions
- Category
- general
- Task
- overall
- Task label
- Overall
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- Artificial Analysis
- Status
- active
- Retrieved at
- 2026-11-07
- Methodology
- Exact AA configuration: Llama 3.1 Nemotron Instruct 70B (release llama-3-1-nemotron-instruct-70b, row llama-3-1-nemotron-instruct-70b, UUID 7393c56a-ec31-48e9-b804-c04f2d2cb641, reasoning=false, effort=not specified, suffix=not specified). This source capture does not provide an average reasoning-token count. Artificial Analysis runs 294 questions from the IFBench test set with five repeats per question. The official evaluator scores prompt-level accuracy in loose mode as pass@1.. Source model ID: artificial-analysis/7393c56a-ec31-48e9-b804-c04f2d2cb641/llama-3-1-nemotron-instruct-70b/llama-3-1-nemotron-instruct-70b. Source model name: Llama 3.1 Nemotron 70B Instruct. Provider reported by source: NVIDIA. Model snapshot: Llama 3.1 Nemotron Instruct 70B. Evaluator: Artificial Analysis. Dataset type: artificial-analysis-evaluation. Tool policy: Single-turn answer evaluated with the official IFBench source code. Attempts: pass@1, 5 repeats per question. Reasoning: non-reasoning. Reasoning effort: none. Score display: 30.7%. Source ID: artificial-analysis. Evidence capture ID: aa-model-page-2026-11-07-b96e3ffc8c5b. Source payload hash: b96e3ffc8c5b052aa56f96d05e89d02b8cc234a75085820fca0ca2a39e498274. Parser version: artificial-analysis-rsc-models-v4. Imported at: 2026-11-07. Source cohort: artificial-analysis-ifbench:AA test set: 294 questions:standalone:non-reasoning:effort=unspecified:suffix=unspecified
- Comparison methodology
- Exact AA configuration: Llama 3.1 Nemotron Instruct 70B (release llama-3-1-nemotron-instruct-70b, row llama-3-1-nemotron-instruct-70b, UUID 7393c56a-ec31-48e9-b804-c04f2d2cb641, reasoning=false, effort=not specified, suffix=not specified). This source capture does not provide an average reasoning-token count. Artificial Analysis runs 294 questions from the IFBench test set with five repeats per question. The official evaluator scores prompt-level accuracy in loose mode as pass@1.. Tool policy: Single-turn answer evaluated with the official IFBench source code. Attempts: pass@1, 5 repeats per question. Reasoning: non-reasoning. Reasoning effort: none
- Context
- The often cited 58 refers to verifiable out-of-domain constraint types, not the number of questions in AA's test run.
Long-RewardBench Pairwise LongQA (Show measurement, test conditions, and source)
- Source value
- 47.1
- Score
- 47.1
- Metric
- Pairwise preference judgment accuracy
- Unit
- %
- Benchmark version
- arXiv:2510.06915v2
- Category
- Long-context reward modeling
- Direction
- Higher is better
- Sample
- 240
- Source type
- independent-evaluator
- Evaluator
- LongRM: Revealing and Unlocking the Context Boundary of Reward Modeling (arXiv:2510.06915v2)
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Long-RewardBench preliminary evaluation, Table 1, Pairwise LongQA column. The model is evaluated as a generative reward/judge model on long-context question, context, and candidate-response inputs. The reported metric is preference-judgment accuracy. The paper identifies the table's shortened “Nemotron-70B-Instruct” row as Llama-3.1-Nemotron-70B-Instruct in its model-selection text. Appendix A reports 240 items for this task column.
- Context
- The paper reports no confidence interval. It does not identify a checkpoint revision, quantization, inference engine, or provider route for this evaluation. This is reward-model/judge performance, not general answer-generation quality.
Long-RewardBench Pairwise Summ (Show measurement, test conditions, and source)
- Source value
- 33
- Score
- 33
- Metric
- Pairwise preference judgment accuracy
- Unit
- %
- Benchmark version
- arXiv:2510.06915v2
- Category
- Long-context reward modeling
- Direction
- Higher is better
- Sample
- 200
- Source type
- independent-evaluator
- Evaluator
- LongRM: Revealing and Unlocking the Context Boundary of Reward Modeling (arXiv:2510.06915v2)
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Long-RewardBench preliminary evaluation, Table 1, Pairwise Summ column (the source's summarization header). The model is evaluated as a generative reward/judge model on long-context question, context, and candidate-response inputs. The reported metric is preference-judgment accuracy. The paper identifies the table's shortened “Nemotron-70B-Instruct” row as Llama-3.1-Nemotron-70B-Instruct in its model-selection text. Appendix A reports 200 items for this task column.
- Context
- The paper reports no confidence interval. It does not identify a checkpoint revision, quantization, inference engine, or provider route for this evaluation. This is reward-model/judge performance, not general answer-generation quality.
Long-RewardBench Pairwise Safety (Show measurement, test conditions, and source)
- Source value
- 57.1
- Score
- 57.1
- Metric
- Pairwise preference judgment accuracy
- Unit
- %
- Benchmark version
- arXiv:2510.06915v2
- Category
- Long-context reward modeling
- Direction
- Higher is better
- Sample
- 140
- Source type
- independent-evaluator
- Evaluator
- LongRM: Revealing and Unlocking the Context Boundary of Reward Modeling (arXiv:2510.06915v2)
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Long-RewardBench preliminary evaluation, Table 1, Pairwise Safety column. The model is evaluated as a generative reward/judge model on long-context question, context, and candidate-response inputs. The reported metric is preference-judgment accuracy. The paper identifies the table's shortened “Nemotron-70B-Instruct” row as Llama-3.1-Nemotron-70B-Instruct in its model-selection text. Appendix A reports 140 items for this task column.
- Context
- The paper reports no confidence interval. It does not identify a checkpoint revision, quantization, inference engine, or provider route for this evaluation. This is reward-model/judge performance, not general answer-generation quality.
Long-RewardBench Pairwise ICL (Show measurement, test conditions, and source)
- Source value
- 41.3
- Score
- 41.3
- Metric
- Pairwise preference judgment accuracy
- Unit
- %
- Benchmark version
- arXiv:2510.06915v2
- Category
- Long-context reward modeling
- Direction
- Higher is better
- Sample
- 160
- Source type
- independent-evaluator
- Evaluator
- LongRM: Revealing and Unlocking the Context Boundary of Reward Modeling (arXiv:2510.06915v2)
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Long-RewardBench preliminary evaluation, Table 1, Pairwise ICL column. The model is evaluated as a generative reward/judge model on long-context question, context, and candidate-response inputs. The reported metric is preference-judgment accuracy. The paper identifies the table's shortened “Nemotron-70B-Instruct” row as Llama-3.1-Nemotron-70B-Instruct in its model-selection text. Appendix A reports 160 items for this task column.
- Context
- The paper reports no confidence interval. It does not identify a checkpoint revision, quantization, inference engine, or provider route for this evaluation. This is reward-model/judge performance, not general answer-generation quality.
Long-RewardBench Pairwise Cite (Show measurement, test conditions, and source)
- Source value
- 54
- Score
- 54
- Metric
- Pairwise preference judgment accuracy
- Unit
- %
- Benchmark version
- arXiv:2510.06915v2
- Category
- Long-context reward modeling
- Direction
- Higher is better
- Sample
- 100
- Source type
- independent-evaluator
- Evaluator
- LongRM: Revealing and Unlocking the Context Boundary of Reward Modeling (arXiv:2510.06915v2)
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Long-RewardBench preliminary evaluation, Table 1, Pairwise Cite column. The model is evaluated as a generative reward/judge model on long-context question, context, and candidate-response inputs. The reported metric is preference-judgment accuracy. The paper identifies the table's shortened “Nemotron-70B-Instruct” row as Llama-3.1-Nemotron-70B-Instruct in its model-selection text. Appendix A reports 100 items for this task column.
- Context
- The paper reports no confidence interval. It does not identify a checkpoint revision, quantization, inference engine, or provider route for this evaluation. This is reward-model/judge performance, not general answer-generation quality.
Long-RewardBench Pairwise Code (Show measurement, test conditions, and source)
- Source value
- 53
- Score
- 53
- Metric
- Pairwise preference judgment accuracy
- Unit
- %
- Benchmark version
- arXiv:2510.06915v2
- Category
- Long-context reward modeling
- Direction
- Higher is better
- Sample
- 100
- Source type
- independent-evaluator
- Evaluator
- LongRM: Revealing and Unlocking the Context Boundary of Reward Modeling (arXiv:2510.06915v2)
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Long-RewardBench preliminary evaluation, Table 1, Pairwise Code column. The model is evaluated as a generative reward/judge model on long-context question, context, and candidate-response inputs. The reported metric is preference-judgment accuracy. The paper identifies the table's shortened “Nemotron-70B-Instruct” row as Llama-3.1-Nemotron-70B-Instruct in its model-selection text. Appendix A reports 100 items for this task column.
- Context
- The paper reports no confidence interval. It does not identify a checkpoint revision, quantization, inference engine, or provider route for this evaluation. This is reward-model/judge performance, not general answer-generation quality.
Long-RewardBench Pairwise Math (Show measurement, test conditions, and source)
- Source value
- 65
- Score
- 65
- Metric
- Pairwise preference judgment accuracy
- Unit
- %
- Benchmark version
- arXiv:2510.06915v2
- Category
- Long-context reward modeling
- Direction
- Higher is better
- Sample
- 60
- Source type
- independent-evaluator
- Evaluator
- LongRM: Revealing and Unlocking the Context Boundary of Reward Modeling (arXiv:2510.06915v2)
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Long-RewardBench preliminary evaluation, Table 1, Pairwise Math column. The model is evaluated as a generative reward/judge model on long-context question, context, and candidate-response inputs. The reported metric is preference-judgment accuracy. The paper identifies the table's shortened “Nemotron-70B-Instruct” row as Llama-3.1-Nemotron-70B-Instruct in its model-selection text. Appendix A reports 60 items for this task column.
- Context
- The paper reports no confidence interval. It does not identify a checkpoint revision, quantization, inference engine, or provider route for this evaluation. This is reward-model/judge performance, not general answer-generation quality.
Long-RewardBench Best-of-N Rank2 (Show measurement, test conditions, and source)
- Source value
- 25.7
- Score
- 25.7
- Metric
- Rank Match Ratio
- Unit
- %
- Benchmark version
- arXiv:2510.06915v2
- Category
- Long-context reward modeling
- Direction
- Higher is better
- Sample
- 300
- Source type
- independent-evaluator
- Evaluator
- LongRM: Revealing and Unlocking the Context Boundary of Reward Modeling (arXiv:2510.06915v2)
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Long-RewardBench preliminary evaluation, Table 1, Best-of-N Rank2 column. The model is evaluated as a generative reward/judge model on long-context question, context, and candidate-response inputs. The score is the Rank Match Ratio, meaning the share of positions where the predicted ranking matches the gold ranking. The paper identifies the table's shortened “Nemotron-70B-Instruct” row as Llama-3.1-Nemotron-70B-Instruct in its model-selection text. Appendix A reports 300 items for this task column.
- Context
- The paper reports no confidence interval. It does not identify a checkpoint revision, quantization, inference engine, or provider route for this evaluation. This is reward-model/judge performance, not general answer-generation quality.
Long-RewardBench Best-of-N Rank3 (Show measurement, test conditions, and source)
- Source value
- 22.3
- Score
- 22.3
- Metric
- Rank Match Ratio
- Unit
- %
- Benchmark version
- arXiv:2510.06915v2
- Category
- Long-context reward modeling
- Direction
- Higher is better
- Sample
- 300
- Source type
- independent-evaluator
- Evaluator
- LongRM: Revealing and Unlocking the Context Boundary of Reward Modeling (arXiv:2510.06915v2)
- Status
- active
- Retrieved at
- 2026-10-17
- Methodology
- Independent evaluation. Long-RewardBench preliminary evaluation, Table 1, Best-of-N Rank3 column. The model is evaluated as a generative reward/judge model on long-context question, context, and candidate-response inputs. The score is the Rank Match Ratio, meaning the share of positions where the predicted ranking matches the gold ranking. The paper identifies the table's shortened “Nemotron-70B-Instruct” row as Llama-3.1-Nemotron-70B-Instruct in its model-selection text. Appendix A reports 300 items for this task column.
- Context
- The paper reports no confidence interval. It does not identify a checkpoint revision, quantization, inference engine, or provider route for this evaluation. This is reward-model/judge performance, not general answer-generation quality.
Page 1 of 4
Popularity compared
How often the model is searched for, used and mentioned, compared with every model of its category.
Popularity index
The 100 most popular models of this category, 0 to 100, higher is more popular
20 of 101 models
As of November 2026 · Source: gradually.ai popularity index
Over time
Weekly values from October 18, 2026 to November 7, 2026
Llama 3.1 Nemotron 70B InstructHolo4-35B-A3B 35MiMo-V2.5-Pro 35Magistral Small 1.2 35Mistral Large 2407 35
Method changes in the timeline, a jump at these points is no market event:
- 1November 1, 2026: Value only with search or usage data, 30-day attention
What is gaining popularity right now, across models, vendors, harnesses and AI subscriptions, is collected in All AI trends.
Similar models
More models of the same kind as Llama 3.1 Nemotron 70B Instruct, chosen by model line, developer, size, price and release date.
- Llama 3.3 Nemotron Super 49B v1NVIDIA · same developer
- Llama 3.3 Nemotron Super 49B v1.5NVIDIA · same developer
- Nemotron 3 SuperNVIDIA · same developer
- Nemotron 3 NanoNVIDIA · same developer
- Nemotron Cascade 2 30B A3BNVIDIA · same developer
- Llama 3.1 Nemotron Ultra 253BNVIDIA · same developer
- Nemotron Nano 2 VL 12BNVIDIA · same developer
- Nemotron Nano 2 9BNVIDIA · same developer
All specifications
Architecture
Input and output
Watermarking and provenance
AI watermarking refers to invisible signals embedded in the output. Visible marks and provenance metadata are listed separately. Unknown means there is insufficient evidence for that particular field.
- AI watermarking
- Unknown
- no source
- Watermarking technique
- Not known
- no source
Licence and access
Frequently asked questions about Llama 3.1 Nemotron 70B Instruct
How large is the context window of Llama 3.1 Nemotron 70B Instruct?
128,000 tokens. That is how much text Llama 3.1 Nemotron 70B Instruct reads at once, about 190 A4 pages.
Page count as a rule of thumb from Artificial Analysis
GLM-5.1200K
DeepSeek-V3 0324164K
GLM-4.5131K
GPT OSS 120B131K
Llama 3.1 Nemotron 70B InstructRank 98 of 103128K
DeepSeek-V3.2128K
DeepSeek-R1128K
Can I run Llama 3.1 Nemotron 70B Instruct locally?
Yes, the weights are open.
Stated by NVIDIA
- Open weights
- yes
- License
- Llama 3.1 Community
- Parameters
- 70B
- active per token
- n/a
Sources and data date
Show all 23 sources
Benchmark sources at a glance
Which independent evaluators tested this model. This is coverage, not a quality ranking.
| Evaluator | Test series | Composite indices | Recorded active measurements |
|---|---|---|---|
| Artificial Analysis | 1 | 0 | 1 |
| LMArena | 0 | 0 | 1 |
Without vendor claims and historical or disputed results; those stay labelled in the single values. 1 measurements are not yet assigned to a test series.