# Grok 3 vs Llama 3-70B

> Grok 3 is the stronger model overall, scoring 39.9 to 28.8 on the Noometry Index.

- Canonical page: https://noometry.com/compare/grok-3-vs-llama-3-70b
- Last updated: 2026-10-11
- Shared benchmarks: 22

## Summary

- They share 22 benchmarks with published results for both. Grok 3 scores higher in 8 categories and Llama 3-70B in 1 category; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Grok 3 leads 46.2 to 20.8.
- The biggest single-benchmark swing is MATH Level 5: 88.7% for Grok 3 and 22.6% for Llama 3-70B.
- Llama 3-70B has downloadable open weights; the other is API-only.

## Snapshot

| | Grok 3 | Llama 3-70B |
|---|---|---|
| Provider | xAI | Meta |
| Noometry Index | 39.9 | 28.8 |
| Rank | 157 | 323 |
| Context | — | — |
| Input $/M | — | — |
| Output $/M | — | — |
| Weights | Proprietary | Open |

## Coding

- Grok 3: 41.9 (#115)
- Llama 3-70B: 35.8 (#218)

| Benchmark | Grok 3 | Llama 3-70B |
|---|---|---|
| LMArena Coding | 1432 | 1206 |
| Aider Polyglot | 53.3% | — |
| WeirdML | 37.2% | — |
| BigCodeBench Instruct | — | 43.6% |
| BigCodeBench Complete | — | 54.5% |
| HumanEval+ | — | 72% |
| MBPP+ | — | 69% |

## Agentic & Tool Use

- Grok 3: 30.5 (#76)
- Llama 3-70B: 21.1 (#139)

| Benchmark | Grok 3 | Llama 3-70B |
|---|---|---|
| Cybench | — | 5% |
| BALROG | 29.5% | — |

## Reasoning

- Grok 3: 13.7 (#333)
- Llama 3-70B: 18.0 (#288)

| Benchmark | Grok 3 | Llama 3-70B |
|---|---|---|
| Kagi LLM Benchmark | 61.3% | 35.1% |
| LMArena Hard Prompts | 1434 | 1195 |
| Epoch Capabilities Index | 138.33 | 122.93 |
| ARC-AGI-2 | 0% | — |
| SimpleBench | 36.1% | — |
| ARC-AGI-1 | 5.5% | — |
| DTBench | — | 54.2% |
| ForecastBench | — | 57.1 |
| WinoGrande | — | 83.5% |

## Math

- Grok 3: 38.0 (#145)
- Llama 3-70B: 12.8 (#305)

| Benchmark | Grok 3 | Llama 3-70B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 55.6% | 4.3% |
| LMArena Math | 1391 | 1218 |
| MATH Level 5 | 88.7% | 22.6% |
| Omni-MATH | 46.4% | — |
| FrontierMath (Feb 2025 set) | 3.8% | — |
| FrontierMath Tier 4 (v1) | 0% | — |

## Knowledge

- Grok 3: 46.2 (#82)
- Llama 3-70B: 20.8 (#277)

| Benchmark | Grok 3 | Llama 3-70B |
|---|---|---|
| GPQA Diamond | 75.8% | 40.6% |
| LMArena Expert | 1421 | 1149 |
| MMLU-Pro | 78.8% | — |
| Confabulations | 14.2% | — |
| Vectara Hallucination Rate | 5.8% | — |
| GPQA (HELM) | 65% | — |
| MMLU | — | 79.3% |

## Multilingual

- Grok 3: 52.3 (#87)
- Llama 3-70B: 33.6 (#251)

| Benchmark | Grok 3 | Llama 3-70B |
|---|---|---|
| LMArena Non-English | 1410 | 1142 |
| LMArena Chinese | 1448 | 1114 |
| LMArena French | 1460 | 1232 |
| LMArena German | 1431 | 1169 |
| LMArena Japanese | 1387 | 1017 |
| LMArena Korean | 1373 | 1017 |
| LMArena Russian | 1416 | 1159 |
| LMArena Spanish | 1417 | 1241 |

## Instruction Following

- Grok 3: 75.0 (#73)
- Llama 3-70B: 62.5 (#238)

| Benchmark | Grok 3 | Llama 3-70B |
|---|---|---|
| LMArena Instruction Following | 1409 | 1194 |
| IFEval | 88.4% | — |

## Long Context

- Grok 3: 38.7 (#192)
- Llama 3-70B: 35.6 (#240)

| Benchmark | Grok 3 | Llama 3-70B |
|---|---|---|
| LMArena Longer Query | 1439 | 1174 |
| Fiction.LiveBench | 58.3% | — |

## Writing & Preference

- Grok 3: 55.8 (#141)
- Llama 3-70B: 42.8 (#231)

| Benchmark | Grok 3 | Llama 3-70B |
|---|---|---|
| LMArena Text | 1426 | 1221 |
| LMArena Creative Writing | 1414 | 1210 |
| LMArena Multi-Turn | 1425 | 1223 |
| Short-Story Creative Writing | 76.4% | — |
| EQ-Bench Creative Writing | 1186 | — |
| WildBench | 84.9% | — |

## FAQ

### Is Grok 3 better than Llama 3-70B?

Grok 3 is the stronger model overall, scoring 39.9 to 28.8 on the Noometry Index.

### Is Grok 3 or Llama 3-70B better for coding?

Grok 3 scores higher on coding benchmarks: 41.9 versus 35.8 in the Noometry coding category.

### How many benchmarks do Grok 3 and Llama 3-70B share?

22 benchmarks have published results for both models. Grok 3 has 40 scored results on Noometry and Llama 3-70B has 31.