Meta Llama 3 8B By meta-llama: Benchmarks, Features and Detailed Analysis. Insights on Meta Llama 3 8B.

Autotrain compatible En Endpoints compatible Facebook Llama Llama-3 Meta Pytorch Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/meta-llama/Meta-Llama-3-8B

Meta Llama 3 8B Benchmarks

MMLU Pro: 24.55

GPQA: 7.38

MUSR: 6.24

BBH: 24.5

IFEval: 14.55 vs 88 (so35)^-83.5%

ARC: 59.22 vs 96.7 (so35)^-38.8%

HellaSwag: 82.02 vs 95.3 (gpt4)^-13.9%

MMLU: 66.49 vs 88.3 (so35)^-24.7%

TruthfulQA: 43.95 vs 59 (gpt4)^-25.5%

WinoGrande: 77.11 vs 87.5 (gpt4)^-11.9%

GSM8K: 45.34 vs 96.4 (so35)^-53%

MATH Lvl 5: 4.53

LLME Score: 0.47253

^nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

What is the LLM Explorer Rank (Score)

Meta Llama 3 8B (meta-llama/Meta-Llama-3-8B)

🌟 Advertise your project 🚀

Meta Llama 3 8B Parameters and Internals

Model Type

text generation, multimodal

Use Cases

Areas:

commercial use, research

Applications:

assistant-like chat, natural language generation

Primary Use Cases:

pretrained models adapted for various NLG tasks

Limitations:

Only tested in English, Not all scenarios addressed.

Considerations:

Developers can fine-tune Llama 3 for languages beyond English, compliance required.

Additional Notes

Some trade-off between model helpfulness and alignment likely unavoidable.

Supported Languages

English (Primary)

Training Details

Data Sources:

publicly available online data

Data Volume:

15 trillion tokens

Methodology:

auto-regressive transformer architecture, supervised fine-tuning, reinforcement learning with human feedback (RLHF)

Context Length:

8000

Training Time:

7.7M GPU hours

Hardware Used:

H100-80GB (TDP of 700W)

Model Architecture:

optimized transformer architecture

Safety Evaluation

Methodologies:

red teaming, adversarial evaluations

Findings:

residual risks may remain

Risk Categories:

child safety risks, security risks, bias risks

Ethical Considerations:

Openness, inclusivity, and helpfulness valued.

Responsible Ai Considerations

Fairness:

Consideration given to bias and fairness during development.

Transparency:

Safety evaluations and risk assessments detailed.

Accountability:

Developers advised to perform application-specific safety testing.

Mitigation Strategies:

Red teaming, adversarial evaluations, safety mitigations included.

Input Output

Input Format:

text

Accepted Modalities:

text

Output Format:

text and code

Release Notes

Version:

3.1-8B

Date:

2024-04-18

Notes:

Released with improved performance for dialogue use cases.

LLM Name	Meta Llama 3 8B
Repository 🤗	https://huggingface.co/meta-llama/Meta-Llama-3-8B
Model Size	8b
Required VRAM	16.1 GB
Updated	2025-10-22
Maintainer	meta-llama
Model Type	llama
Model Files	5.0 GB: 1-of-4 5.0 GB: 2-of-4 4.9 GB: 3-of-4 1.2 GB: 4-of-4
Supported Languages	en
Model Architecture	LlamaForCausalLM
License	llama3
Context Length	8192
Model Max Length	8192
Transformers Version	4.40.0.dev0
Tokenizer Class	PreTrainedTokenizerFast
Vocabulary Size	128256
Torch Data Type	bfloat16

Quantized Models of the Meta Llama 3 8B

Model	Likes	Downloads	VRAM
Llama 3 8B Bnb 4bit	202	37632	5 GB
Monah 8B Uncensored V0.2 Gguf	5	138	3 GB
Monah 8B Uncensored Gguf	4	120	3 GB
Meta Llama 3 8B AWQ	0	1023	5 GB
Llama 3 8B GPTQ 4 Bit	6	989	5 GB
Llama3 8B Chips 4bit	0	6	5 GB
Monah 8B Uncensored V0.2 Gguf	5	131	3 GB
Hyper X Llama 3 Small Python	0	5	5 GB
Openchat 3.6 8B 20240522 AWQ	1	8	5 GB
...at 3.6 8B 20240522 8 0bpw EXL2	2	8	8 GB

Best Alternatives to Meta Llama 3 8B

Best Alternatives	Context / RAM	Downloads	Likes
...otron 8B UltraLong 4M Instruct	4192K / 32.1 GB	1531	120
...a 3.1 8B UltraLong 4M Instruct	4192K / 32.1 GB	176	24
UltraLong Thinking	4192K / 16.1 GB	45	3
...otron 8B UltraLong 2M Instruct	2096K / 32.1 GB	1142	15
...a 3.1 8B UltraLong 2M Instruct	2096K / 32.1 GB	875	9
...otron 8B UltraLong 1M Instruct	1048K / 32.1 GB	7142	52
...a 3.1 8B UltraLong 1M Instruct	1048K / 32.1 GB	1387	29
...xis Bookwriter Llama3.1 8B Sft	1048K / 16.1 GB	40	4
Zero Llama 3.1 8B Beta6	1048K / 16.1 GB	2	1
...dger Nu Llama 3.1 8B UltraLong	1048K / 16.2 GB	5	3

Rank the Meta Llama 3 8B Capabilities

🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

What open-source LLMs or SLMs are you in search of? 51544 in total.

Email us: info@extractum.io. Our Privacy Policy | Terms and Conditions | Suggest an improvement.

Our Social Media →

Original data from HuggingFace, OpenCompass and various public git repos.

Release v20241124

Support LLM Explorer