Rankings / NVIDIA
Llama 3.1 Nemotron Ultra 253B V1
open weightsreleased 2025-04-07
BenchAtlas Index
as of 2026-07-14
thinking
rank #372
10 families · 5 categories · high
Benchmark evidence
23 results
agentic coding
| τ²-Bench | thinking | 11.4 % | independent | Artificial Analysis |
coding
| LiveCodeBench v5older version problems=279 · window_start=2024-08-01 | 64.5 % | independent | LiveCodeBench Leaderboard | |
| LiveCodeBench v6 implementation=artificial-analysis | thinking | 64.1 % | independent | Artificial Analysis |
| SciCode | thinking | 34.7 % | independent | Artificial Analysis |
external indices
| Intelligence Index v4.1 | thinking | 9.1 points | independent | Artificial Analysis |
| AA Math Index | thinking | 63.7 points | independent | Artificial Analysis |
human preference
| Industry Software And It Services (style control)older version arena=text · category=industry_software_and_it_services · style_control=true | 1387.4 1363.8–1411.0 | community | LMArena Leaderboard Dataset | |
| English (style control)older version arena=text · category=english · style_control=true | 1375.6 1360.7–1390.5 | community | LMArena Leaderboard Dataset | |
| Hard Prompts (style control)older version arena=text · category=hard_prompts · style_control=true | 1374.7 1353.4–1395.9 | community | LMArena Leaderboard Dataset | |
| Instruction Following (style control)older version arena=text · category=instruction_following · style_control=true | 1351.3 1329.6–1372.9 | community | LMArena Leaderboard Dataset | |
| Overall (style control) arena=text · category=overall · style_control=true | 1347.5 1335.8–1359.2 | community | LMArena Leaderboard Dataset | |
| Industry Writing And Literature And Language (style control)older version arena=text · category=industry_writing_and_literature_and_language · style_control=true | 1320.4 1299.4–1341.3 | community | LMArena Leaderboard Dataset | |
| Non English (style control)older version arena=text · category=non_english · style_control=true | 1305.9 1288.9–1323.0 | community | LMArena Leaderboard Dataset | |
| Exclude Ties (style control)older version arena=text · category=exclude_ties · style_control=true | 1303.9 1286.3–1321.5 | community | LMArena Leaderboard Dataset | |
| Industry Entertainment And Sports And Media (style control)older version arena=text · category=industry_entertainment_and_sports_and_media · style_control=true | 1303.0 1279.2–1326.8 | community | LMArena Leaderboard Dataset |
knowledge science
| GPQA Diamond implementation=artificial-analysis | thinking | 72.8 % | independent | Artificial Analysis |
| Humanity's Last Exam implementation=artificial-analysis | thinking | 8.1 % | independent | Artificial Analysis |
| MMLU-Pro implementation=artificial-analysis | thinking | 82.5 % | independent | Artificial Analysis |
long context instruction
| AA-LCR | thinking | 7.3 % | independent | Artificial Analysis |
| IFBench | thinking | 38.2 % | independent | Artificial Analysis |
reasoning math
| AIME implementation=artificial-analysis | thinking | 74.7 % | independent | Artificial Analysis |
| AIME year=2025 · implementation=artificial-analysis | thinking | 63.7 % | independent | Artificial Analysis |
| MATH-500 implementation=artificial-analysis | thinking | 95.2 % | independent | Artificial Analysis |
Agent + model results
systems, not bare-model scores
| agent + model Artificial Analysis harness + Llama 3.1 Nemotron Ultra 253B V1 | Terminal-Bench Hard | 2.3 % | independent | Artificial Analysis |
These scores measure the whole agent system (scaffold, tools, budgets) — they are never merged into the bare model’s numbers.