# GPT-4 vs Llama 3.1-405B

> Llama 3.1-405B is the stronger model overall, scoring 30.7 to 29.1 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gpt-4-vs-llama-3-1-405b
- Last updated: 2026-10-12
- Shared benchmarks: 30

## Summary

- They share 30 benchmarks with published results for both. GPT-4 scores higher in 1 category and Llama 3.1-405B in 7 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Llama 3.1-405B leads 30.4 to 18.4.
- The biggest single-benchmark swing is MATH Level 5: 23% for GPT-4 and 49.8% for Llama 3.1-405B.
- Llama 3.1-405B has downloadable open weights; the other is API-only.

## Snapshot

| | GPT-4 | Llama 3.1-405B |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 29.1 | 30.7 |
| Rank | 316 | 288 |
| Context | 8K | — |
| Input $/M | $30 | — |
| Output $/M | $60 | — |
| Weights | Proprietary | Open |

## Coding

- GPT-4: 31.6 (#283)
- Llama 3.1-405B: 33.1 (#262)

| Benchmark | GPT-4 | Llama 3.1-405B |
|---|---|---|
| WeirdML | 12.4% | 21.4% |
| LMArena Coding | 1254 | 1291 |
| BigCodeBench Instruct | 46% | — |
| BigCodeBench Complete | 57.2% | — |
| HumanEval+ | 79.3% | — |

## Agentic & Tool Use

- GPT-4: —
- Llama 3.1-405B: 21.0 (#140)

| Benchmark | GPT-4 | Llama 3.1-405B |
|---|---|---|
| TheAgentCompany | — | 7.4% |
| Cybench | — | 7.5% |
| METR Time Horizons | 36.1% | — |

## Reasoning

- GPT-4: 17.8 (#289)
- Llama 3.1-405B: 16.8 (#300)

| Benchmark | GPT-4 | Llama 3.1-405B |
|---|---|---|
| LMArena Hard Prompts | 1241 | 1269 |
| DTBench | 62.7% | 61.4% |
| BIG-Bench Hard | 75.1% | 82.9% |
| Epoch Capabilities Index | 125.89 | 128.75 |
| ForecastBench | 57.8 | 59.9 |
| HellaSwag | 95.3% | 89.2% |
| WinoGrande | 87.5% | 89.2% |
| SimpleBench | — | 23% |
| Kagi LLM Benchmark | — | 45% |
| Chess Puzzles | 4% | — |
| Mystery Game Puzzles | 12% | — |
| LMCA | 17.1% | — |
| PIQA | — | 85.9% |

## Math

- GPT-4: 10.8 (#309)
- Llama 3.1-405B: 18.4 (#290)

| Benchmark | GPT-4 | Llama 3.1-405B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.1% | 9.7% |
| LMArena Math | 1269 | 1281 |
| MATH Level 5 | 23% | 49.8% |
| Omni-MATH | — | 24.9% |
| GSM8K | 92% | — |

## Knowledge

- GPT-4: 18.4 (#282)
- Llama 3.1-405B: 30.4 (#227)

| Benchmark | GPT-4 | Llama 3.1-405B |
|---|---|---|
| GPQA Diamond | 35.7% | 50.9% |
| LMArena Expert | 1211 | 1243 |
| MMLU | 86.4% | 84.5% |
| TriviaQA | 84.8% | 82.7% |
| MMLU-Pro | — | 72.3% |
| Confabulations | — | 17.6% |
| GPQA (HELM) | — | 52.2% |
| ARC (AI2) Challenge | — | 95.3% |

## Multilingual

- GPT-4: 40.6 (#215)
- Llama 3.1-405B: 40.7 (#214)

| Benchmark | GPT-4 | Llama 3.1-405B |
|---|---|---|
| LMArena Non-English | 1246 | 1248 |
| LMArena Chinese | 1242 | 1242 |
| LMArena French | 1283 | 1279 |
| LMArena German | 1251 | 1252 |
| LMArena Japanese | 1209 | 1208 |
| LMArena Korean | 1184 | 1184 |
| LMArena Russian | 1251 | 1265 |
| LMArena Spanish | 1261 | 1260 |

## Instruction Following

- GPT-4: 65.3 (#222)
- Llama 3.1-405B: 65.9 (#214)

| Benchmark | GPT-4 | Llama 3.1-405B |
|---|---|---|
| LMArena Instruction Following | 1241 | 1259 |
| IFEval | — | 81.1% |

## Long Context

- GPT-4: 37.7 (#212)
- Llama 3.1-405B: 38.4 (#197)

| Benchmark | GPT-4 | Llama 3.1-405B |
|---|---|---|
| LMArena Longer Query | 1244 | 1266 |

## Writing & Preference

- GPT-4: 34.9 (#268)
- Llama 3.1-405B: 38.9 (#251)

| Benchmark | GPT-4 | Llama 3.1-405B |
|---|---|---|
| LMArena Text | 1263 | 1284 |
| LMArena Creative Writing | 1244 | 1262 |
| EQ-Bench Creative Writing | 752 | 870 |
| LMArena Multi-Turn | 1257 | 1297 |
| WildBench | — | 78.3% |

## FAQ

### Is GPT-4 better than Llama 3.1-405B?

Llama 3.1-405B is the stronger model overall, scoring 30.7 to 29.1 on the Noometry Index.

### Is GPT-4 or Llama 3.1-405B better for coding?

Llama 3.1-405B scores higher on coding benchmarks: 33.1 versus 31.6 in the Noometry coding category.

### How many benchmarks do GPT-4 and Llama 3.1-405B share?

30 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and Llama 3.1-405B has 42.