NVIDIA model product
A large language model customized by NVIDIA to improve the helpfulness of LLM generated responses. It is a fine-tuned version of Llama 3.1 70B Instruct. The model was trained using RLHF (REINFORCE) with HelpSteer2-Preference prompts.
Updated Aug 10, 2026. Default version: Llama 3.1 Nemotron 70B Instruct
Structured fields from the published default version.
This profile uses the latest version under this unique model that has a calculated LLMBoard score. Arena and price are excluded.
Published benchmark records for the scored version Llama 3.1 Nemotron 70B Instruct.
| GSM8K Chat | 0.8 | 1 | 1 | 100.0% | C | |
| Instruct HumanEval | 0.7 | 1 | 1 | 100.0% | C | |
| MMLU Chat | 0.8 | 1 | 1 | 100.0% | C | |
| XLSum English | 0.3 | 1 | 1 | 100.0% | C | |
| Winogrande | 0.8 | 5 | 22 | 81.0% | C | |
| TruthfulQA | 0.6 | 7 | 18 | 64.7% | C | |
| HellaSwag | 0.9 | 12 | 27 | 57.7% | C | |
| MT-Bench | 0.1 | 12 | 12 | 0.0% | C | |
| ARC-C | 0.7 | 23 | 34 | 33.3% | C | |
| GSM8k | 0.9 | 23 | 48 | 53.2% | C | |
| MMLU | 0.8 | 61 | 100 | 39.4% | C |
Preference and agent-evaluation signals from published Arena datasets.
| text | industry medicine and healthcare | 212 | 1300.4 | 392 | N/A | |
| text | industry legal and government | 216 | 1303.1 | 489 | N/A | |
| text | industry mathematical | 218 | 1287.3 | 874 | N/A | |
| text | english | 221 | 1315.1 | 3,714 | N/A | |
| text | creative writing | 223 | 1268.9 | 1,001 | N/A | |
| text style control | industry medicine and healthcare | 224 | 1335.4 | 392 | N/A | |
| text | chinese | 229 | 1262.0 | 723 | N/A | |
| text | industry business and management and financial operations | 229 | 1261.4 | 904 | N/A | |
| text | expert | 230 | 1241.6 | 453 | N/A | |
| text | industry life and physical and social science | 231 | 1293.5 | 1,299 | N/A | |
| text style control | industry mathematical | 231 | 1300.1 | 874 | N/A | |
| text | industry entertainment and sports and media | 233 | 1249.4 | 961 | N/A | |
| text | math | 234 | 1271.1 | 1,041 | N/A | |
| text | overall | 235 | 1282.8 | 7,140 | N/A | |
| text | exclude ties | 235 | 1211.5 | 4,748 | N/A | |
| text style control | industry legal and government | 235 | 1327.0 | 489 | N/A | |
| text | hard prompts english | 236 | 1284.5 | 1,136 | N/A | |
| text | multi turn | 237 | 1275.4 | 1,227 | N/A | |
| text style control | expert | 237 | 1293.9 | 453 | N/A | |
| text style control | creative writing | 241 | 1276.7 | 1,001 | N/A | |
| text style control | english | 241 | 1329.4 | 3,714 | N/A | |
| text | hard prompts | 242 | 1266.7 | 1,953 | N/A | |
| text | industry software and it services | 242 | 1284.3 | 2,081 | N/A | |
| text | instruction following | 242 | 1252.2 | 2,959 | N/A | |
| text style control | chinese | 242 | 1303.5 | 723 | N/A | |
| text style control | math | 245 | 1278.7 | 1,041 | N/A | |
| text | coding | 247 | 1272.3 | 1,312 | N/A | |
| text style control | hard prompts english | 247 | 1327.3 | 1,136 | N/A | |
| text style control | multi turn | 248 | 1293.1 | 1,227 | N/A | |
| text style control | industry entertainment and sports and media | 249 | 1268.5 | 961 | N/A | |
| text style control | industry business and management and financial operations | 250 | 1294.7 | 904 | N/A | |
| text | non english | 251 | 1244.7 | 3,426 | N/A | |
| text | industry writing and literature and language | 254 | 1252.6 | 1,828 | N/A | |
| text style control | hard prompts | 254 | 1310.2 | 1,953 | N/A | |
| text style control | industry life and physical and social science | 254 | 1317.2 | 1,299 | N/A | |
| text style control | instruction following | 257 | 1280.2 | 2,959 | N/A | |
| text style control | overall | 258 | 1299.0 | 7,140 | N/A | |
| text style control | industry software and it services | 258 | 1324.8 | 2,081 | N/A | |
| text style control | exclude ties | 259 | 1233.3 | 4,748 | N/A | |
| text | russian | 261 | 1226.0 | 1,069 | N/A | |
| text style control | coding | 261 | 1328.4 | 1,312 | N/A | |
| text | longer query | 263 | 1238.7 | 1,049 | N/A | |
| text style control | non english | 266 | 1262.0 | 3,426 | N/A | |
| text style control | industry writing and literature and language | 271 | 1265.3 | 1,828 | N/A | |
| text style control | longer query | 271 | 1280.8 | 1,049 | N/A | |
| text style control | russian | 280 | 1241.3 | 1,069 | N/A |
Official vendor API PAYG pricing is summarized first. The table then lists individual provider offerings without treating their minimum as the official price.
| Nvidia | nvidia/llama-3.1-nemotron-70b-instruct | global | N/A | N/A | 128K |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
All published versions linked to this unique model. The score columns identify the version used by the current overall ranking.
| Llama 3.1 Nemotron 70B Instruct | 43.8 | 70B | 128K | 8.2K | Yes | Llama 3.1 Community License |
A concise description based on the published model registry.
A large language model customized by NVIDIA to improve the helpfulness of LLM generated responses. It is a fine-tuned version of Llama 3.1 70B Instruct. The model was trained using RLHF (REINFORCE) with HelpSteer2-Preference prompts.
Use the benchmark, Arena and pricing sections above as separate evidence. A missing field means the current data snapshot does not support that claim.
Data snapshot: 2026-08-07. Editorial model content is not available in the backend.
Open a comparison with the three ranked models immediately above and below this model.
Recommendations prioritize the same model type and family, then the closest published LLMBoard score.
Common questions about Llama 3.1 Nemotron 70B.
Llama 3.1 Nemotron 70B's default version was released on Oct 1, 2024.
No official standard PAYG price is currently available for Llama 3.1 Nemotron 70B.
Llama 3.1 Nemotron 70B is published under NVIDIA in the model registry.
The default version has a 128K token context window.
Yes. The default version is marked as open weight under Llama 3.1 Community License.
1 published provider offerings are linked to the default version.
Nearby ranked alternatives include Llama 3.1 Nemotron Nano 8B, DeepSeek-R1, Qwen2.5 14B.