# Grok-2 (Dec 2024) vs Llama 3-8B

> Grok-2 (Dec 2024) is the stronger model overall, scoring 33.4 to 25.3 on the Noometry Index.

- Canonical page: https://noometry.com/compare/grok-2-vs-llama-3-8b
- Last updated: 2026-10-10
- Shared benchmarks: 22

## Summary

- They share 22 benchmarks with published results for both. Grok-2 (Dec 2024) scores higher in 8 categories and Llama 3-8B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Grok-2 (Dec 2024) leads 29.8 to 7.8.
- The biggest single-benchmark swing is MATH Level 5: 63.5% for Grok-2 (Dec 2024) and 6.1% for Llama 3-8B.
- Llama 3-8B has downloadable open weights; the other is API-only.

## Snapshot

| | Grok-2 (Dec 2024) | Llama 3-8B |
|---|---|---|
| Provider | xAI | Meta |
| Noometry Index | 33.4 | 25.3 |
| Rank | 233 | 338 |
| Context | — | 8K |
| Input $/M | — | $0.14 |
| Output $/M | — | $0.14 |
| Weights | Proprietary | Open |

## Coding

- Grok-2 (Dec 2024): 32.9 (#254)
- Llama 3-8B: 30.5 (#285)

| Benchmark | Grok-2 (Dec 2024) | Llama 3-8B |
|---|---|---|
| LMArena Coding | 1288 | 1152 |
| WeirdML | 22.2% | — |
| BigCodeBench Instruct | — | 31.9% |
| LiveBench Coding | 46.4% | — |
| BigCodeBench Complete | — | 36.9% |
| HumanEval+ | — | 56.7% |
| MBPP+ | — | 54.8% |

## Reasoning

- Grok-2 (Dec 2024): 16.4 (#294)
- Llama 3-8B: 13.8 (#321)

| Benchmark | Grok-2 (Dec 2024) | Llama 3-8B |
|---|---|---|
| LMArena Hard Prompts | 1272 | 1133 |
| DTBench | 65.2% | 43.9% |
| Epoch Capabilities Index | 130.48 | 116.33 |
| SimpleBench | 22.7% | — |
| Chess Puzzles | — | 0% |
| LiveBench Reasoning | 54.8% | — |
| LiveBench Data Analysis | 54.5% | — |
| Adversarial NLI | — | 57.3% |
| ForecastBench | — | 58.6 |
| LiveBench | 54.3% | — |
| WinoGrande | — | 75.7% |

## Math

- Grok-2 (Dec 2024): 19.6 (#278)
- Llama 3-8B: 8.3 (#317)

| Benchmark | Grok-2 (Dec 2024) | Llama 3-8B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 11.5% | 1.9% |
| LMArena Math | 1283 | 1151 |
| MATH Level 5 | 63.5% | 6.1% |
| LiveBench Math | 54.9% | — |
| FrontierMath (Feb 2025 set) | 0.7% | — |

## Knowledge

- Grok-2 (Dec 2024): 29.8 (#225)
- Llama 3-8B: 7.8 (#303)

| Benchmark | Grok-2 (Dec 2024) | Llama 3-8B |
|---|---|---|
| GPQA Diamond | 53.8% | 26.1% |
| LMArena Expert | 1254 | 1113 |
| Confabulations | 20.1% | — |
| ARC (AI2) Challenge | — | 82.8% |
| MMLU | — | 68.8% |
| OpenBookQA | — | 82.6% |
| TriviaQA | — | 67.7% |

## Multilingual

- Grok-2 (Dec 2024): 43.3 (#183)
- Llama 3-8B: 31.0 (#256)

| Benchmark | Grok-2 (Dec 2024) | Llama 3-8B |
|---|---|---|
| LMArena Non-English | 1282 | 1098 |
| LMArena Chinese | 1289 | 1076 |
| LMArena French | 1317 | 1158 |
| LMArena German | 1287 | 1104 |
| LMArena Japanese | 1244 | 967 |
| LMArena Korean | 1237 | 1003 |
| LMArena Russian | 1286 | 1109 |
| LMArena Spanish | 1281 | 1172 |

## Instruction Following

- Grok-2 (Dec 2024): 67.1 (#197)
- Llama 3-8B: 58.7 (#255)

| Benchmark | Grok-2 (Dec 2024) | Llama 3-8B |
|---|---|---|
| LMArena Instruction Following | 1270 | 1127 |
| LiveBench Instruction Following | 69.6% | — |

## Long Context

- Grok-2 (Dec 2024): 38.7 (#186)
- Llama 3-8B: 34.2 (#246)

| Benchmark | Grok-2 (Dec 2024) | Llama 3-8B |
|---|---|---|
| LMArena Longer Query | 1276 | 1128 |

## Writing & Preference

- Grok-2 (Dec 2024): 48.8 (#192)
- Llama 3-8B: 37.7 (#251)

| Benchmark | Grok-2 (Dec 2024) | Llama 3-8B |
|---|---|---|
| LMArena Text | 1305 | 1166 |
| LMArena Creative Writing | 1284 | 1150 |
| LMArena Multi-Turn | 1290 | 1152 |
| Short-Story Creative Writing | 63.6% | — |
| LiveBench Language | 45.6% | — |

## FAQ

### Is Grok-2 (Dec 2024) better than Llama 3-8B?

Grok-2 (Dec 2024) is the stronger model overall, scoring 33.4 to 25.3 on the Noometry Index.

### Is Grok-2 (Dec 2024) or Llama 3-8B better for coding?

Grok-2 (Dec 2024) scores higher on coding benchmarks: 32.9 versus 30.5 in the Noometry coding category.

### How many benchmarks do Grok-2 (Dec 2024) and Llama 3-8B share?

22 benchmarks have published results for both models. Grok-2 (Dec 2024) has 34 scored results on Noometry and Llama 3-8B has 34.