# GPT-4 vs Nemotron 3 Ultra

> Nemotron 3 Ultra is the stronger model overall, scoring 42.5 to 29.1 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gpt-4-vs-nemotron-3-ultra
- Last updated: 2026-10-10
- Shared benchmarks: 25

## Summary

- They share 25 benchmarks with published results for both. GPT-4 scores higher in 0 categories and Nemotron 3 Ultra in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Nemotron 3 Ultra leads 52.5 to 18.4.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 1.1% for GPT-4 and 86.7% for Nemotron 3 Ultra.
- Nemotron 3 Ultra is cheaper at $0.50 / $2.20 per million input/output tokens, against $30 / $60 for GPT-4.
- Nemotron 3 Ultra accepts more context: 262K tokens versus 8K.
- Nemotron 3 Ultra has downloadable open weights; the other is API-only.

## Snapshot

| | GPT-4 | Nemotron 3 Ultra |
|---|---|---|
| Provider | OpenAI | NVIDIA |
| Noometry Index | 29.1 | 42.5 |
| Rank | 316 | 113 |
| Context | 8K | 262K |
| Input $/M | $30 | $0.50 |
| Output $/M | $60 | $2.20 |
| Weights | Proprietary | Open |

## Coding

- GPT-4: 31.6 (#283)
- Nemotron 3 Ultra: 38.1 (#182)

| Benchmark | GPT-4 | Nemotron 3 Ultra |
|---|---|---|
| WeirdML | 12.4% | 43.5% |
| LMArena Coding | 1254 | 1468 |
| FrontierCode | — | 13.6% |
| SciCode | — | 40.3% |
| BigCodeBench Instruct | 46% | — |
| BigCodeBench Complete | 57.2% | — |
| HumanEval+ | 79.3% | — |

## Agentic & Tool Use

- GPT-4: —
- Nemotron 3 Ultra: 23.4 (#126)

| Benchmark | GPT-4 | Nemotron 3 Ultra |
|---|---|---|
| APEX-Agents | — | 22.7% |
| METR Time Horizons | 36.1% | — |

## Reasoning

- GPT-4: 17.8 (#289)
- Nemotron 3 Ultra: 31.0 (#82)

| Benchmark | GPT-4 | Nemotron 3 Ultra |
|---|---|---|
| Chess Puzzles | 4% | 12% |
| LMArena Hard Prompts | 1241 | 1452 |
| Mystery Game Puzzles | 12% | 20% |
| DTBench | 62.7% | 90.1% |
| LMCA | 17.1% | 36.9% |
| Epoch Capabilities Index | 125.89 | 146.17 |
| CritPt | — | 3.1% |
| BIG-Bench Hard | 75.1% | — |
| ForecastBench | 57.8 | — |
| HellaSwag | 95.3% | — |
| WinoGrande | 87.5% | — |

## Math

- GPT-4: 10.8 (#309)
- Nemotron 3 Ultra: 35.3 (#187)

| Benchmark | GPT-4 | Nemotron 3 Ultra |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.1% | 86.7% |
| LMArena Math | 1269 | 1457 |
| ProofBench | — | 2% |
| MATH Level 5 | 23% | — |
| GSM8K | 92% | — |

## Knowledge

- GPT-4: 18.4 (#282)
- Nemotron 3 Ultra: 52.5 (#61)

| Benchmark | GPT-4 | Nemotron 3 Ultra |
|---|---|---|
| GPQA Diamond | 35.7% | 85.4% |
| LMArena Expert | 1211 | 1472 |
| MMLU | 86.4% | — |
| TriviaQA | 84.8% | — |

## Multilingual

- GPT-4: 40.6 (#215)
- Nemotron 3 Ultra: 53.5 (#64)

| Benchmark | GPT-4 | Nemotron 3 Ultra |
|---|---|---|
| LMArena Non-English | 1246 | 1427 |
| LMArena Chinese | 1242 | 1497 |
| LMArena French | 1283 | 1472 |
| LMArena German | 1251 | 1471 |
| LMArena Korean | 1184 | 1386 |
| LMArena Russian | 1251 | 1417 |
| LMArena Spanish | 1261 | 1454 |
| LMArena Japanese | 1209 | — |

## Instruction Following

- GPT-4: 65.3 (#222)
- Nemotron 3 Ultra: 74.7 (#88)

| Benchmark | GPT-4 | Nemotron 3 Ultra |
|---|---|---|
| LMArena Instruction Following | 1241 | 1418 |

## Long Context

- GPT-4: 37.7 (#212)
- Nemotron 3 Ultra: 43.9 (#81)

| Benchmark | GPT-4 | Nemotron 3 Ultra |
|---|---|---|
| LMArena Longer Query | 1244 | 1437 |

## Writing & Preference

- GPT-4: 34.9 (#268)
- Nemotron 3 Ultra: 66.3 (#36)

| Benchmark | GPT-4 | Nemotron 3 Ultra |
|---|---|---|
| LMArena Text | 1263 | 1445 |
| LMArena Creative Writing | 1244 | 1398 |
| EQ-Bench Creative Writing | 752 | 1692 |
| LMArena Multi-Turn | 1257 | 1411 |

## FAQ

### Is GPT-4 better than Nemotron 3 Ultra?

Nemotron 3 Ultra is the stronger model overall, scoring 42.5 to 29.1 on the Noometry Index.

### Which is cheaper, GPT-4 or Nemotron 3 Ultra?

Nemotron 3 Ultra is cheaper. It lists at $0.50 per million input tokens and $2.20 per million output tokens; GPT-4 lists at $30 and $60.

### Is GPT-4 or Nemotron 3 Ultra better for coding?

Nemotron 3 Ultra scores higher on coding benchmarks: 38.1 versus 31.6 in the Noometry coding category.

### Which has the bigger context window?

Nemotron 3 Ultra does, with 262K tokens against 8K.

### How many benchmarks do GPT-4 and Nemotron 3 Ultra share?

25 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and Nemotron 3 Ultra has 30.