# Llama 3.1 8B Benchmark Scores & Performance

> Llama 3.1 8B by Meta has 7 source-displayable benchmark rows but no public overall score. It remains unranked.

Canonical page: https://benchlm.ai/models/llama-3-1-8b

Last updated: October 14, 2026

## Model Details

| Property | Value |
|----------|-------|
| Creator | Meta |
| Source Type | Open Weight |
| Reasoning Type | Non-Reasoning |
| Context Window | 128K |
| Official model card | [Llama 3.1 8B Instruct model card](https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct) |
| Overall Score | Not computed (unranked) |
| Overall Rank | Unranked |

## Family & Coverage

- Family: Llama 3.1 8B
- Variant: base
- Benchmarks covered: 7 of 628
- Coverage note: Published evidence is shown below; no public overall score is assigned.

## Coding Benchmarks

| Benchmark | Score | Source |
|-----------|-------|--------|
| [HumanEval](/benchmarks/humaneval) | 72.6% | [Meta: Llama 3.1 8B Instruct model card](https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct) |

## Knowledge Benchmarks

| Benchmark | Score | Source |
|-----------|-------|--------|
| [MMLU](/benchmarks/mmlu) | 73% | [Meta: Llama 3.1 8B Instruct model card](https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct) |
| [MMLU-Pro](/benchmarks/mmlu-pro) | 48.3% | [Meta: Llama 3.1 8B Instruct model card](https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct) |
| [GPQA](/benchmarks/gpqa) | 30.4% | [Meta: Llama 3.1 8B Instruct model card](https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct) |

## Instruction Following Benchmarks

| Benchmark | Score | Source |
|-----------|-------|--------|
| [IFEval](/benchmarks/ifeval) | 80.4% | [Meta: Llama 3.1 8B Instruct model card](https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct) |

## Multilingual Benchmarks

| Benchmark | Score | Source |
|-----------|-------|--------|
| [MGSM](/benchmarks/mgsm) | 68.9% | [Meta: Llama 3.1 8B Instruct model card](https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct) |

## Mathematics Benchmarks

| Benchmark | Score | Source |
|-----------|-------|--------|
| [GSM8K](/benchmarks/gsm8k) | 84.5% | [Meta: Llama 3.1 8B Instruct model card](https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct) |

## Other Meta Models

- [Muse Wonder Full](/models/muse-wonder-full) - Score: 74.1
- [Muse Wonder](/models/muse-wonder) - Score: 70.74
- [Muse Spark 1.1](/models/muse-spark-1-1) - Score: 65.59
- [Muse Spark 1.2](/models/muse-spark-1-2) - Score: 65.4
- [Muse Spark](/models/muse-spark) - Score: 59.96
- [Muse Glimmer 30B](/models/muse-glimmer-30b) - Score: 41.3
- [Llama 4 Scout](/models/llama-4-scout) - Score: 29.1
- [Llama 3.1 405B](/models/llama-3-1-405b) - Score: 28.09
- [Llama 4 Maverick](/models/llama-4-maverick) - Score: 21.94
- [Llama 3 70B](/models/llama-3-70b) - Score: 21.27
- [Muse Spark 1.3](/models/muse-spark-1-3) - Score: not computed
- [Llama 3.3 70B](/models/llama-3-3-70b) - Score: not computed
- [Llama 3.1 70B](/models/llama-3-1-70b) - Score: not computed
- [Llama 4 Behemoth](/models/llama-4-behemoth) - Score: not computed
- [Muse Voice Transcribe](/models/muse-voice-transcribe) - Score: not computed