Llama-3.3-Nemotron-Super-49B-v1.5
Llama-3.3-Nemotron-Super-49B-v1.5.
15
Benchmarks with a score
0
Ranked first, out of five or more models
0
In the top three
NVIDIA Jul 2025 Llama-3.3-Nemotron-Super-49B-v1.5
Open weights Mid-size
137.2
#205 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
- LiveCodeBench v5 73.6% #4 /47
- SnorkelSequences 66.8% #9 /25
- FinQA (Snorkel OpenEnv environment) 44% #19 /22
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Knowledge and language
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Arena-Hard-Auto | 92% | #6 of 39 | Lab model card | |
| IFEval | 88.6% | #48 of 135 | Lab model card | |
| MMLU-Pro | 79.5% | #83 of 167 | Lab model card | |
| LMArena (Elo) | 1,343 | #185 of 335 | Official leaderboard | |
| Humanity's Last Exam | 7.64% | #146 of 154 | Lab model card |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MATH-500 | 97.4% | #15 of 70 | Lab model card | |
| AIME 2024 | 87.5% | #25 of 88 | Lab model card | |
| SnorkelSequences | 66.8% | #9 of 25 | Official leaderboard | |
| GPQA Diamond | 72% | #166 of 302 | Lab model card | |
| AIME 2025 | 82.7% | #84 of 143 | Lab model card |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| LiveCodeBench v5 | 73.6% | #4 of 47 | Lab model card |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Berkeley Function Calling Leaderboard (BFCL) | 71.8% | #13 of 41 | Lab model card |
Professional work
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Finance Reasoning (Snorkel) | 35.4% | #15 of 21 | Official leaderboard | |
| FinQA (Snorkel OpenEnv environment) | 44% | #19 of 22 | Official leaderboard | |
| SnorkelFinance | 44% | #19 of 22 | Official leaderboard |
Its lab's numbers put it at 137.2, number 205 of 404
The Self-Reported Index fits one score to the 10 capability results NVIDIA reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 134.7 to 139.7. How we compute it.
- MATH-500 97.4
- Arena-Hard-Auto 92
- IFEval 88.6
- AIME 2024 87.5
- AIME 2025 82.7
- MMLU-Pro 79.5
- LiveCodeBench v5 73.6
- GPQA Diamond 72
- Berkeley Function Calling Leaderboard (BFCL) 71.8
- Humanity's Last Exam 7.64
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.