What are the hardware requirements for Phi 3 Mini 128K Instruct?

Phi 3 Mini 128K Instruct requires approximately 7.7 GB of VRAM and supports a context window of 128K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Phi 3 Mini 128K Instruct and how large is it?

Phi 3 Mini 128K Instruct is developed by microsoft, a model with 3.8b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Phi 3 Mini 128K Instruct perform on standard benchmarks?

Phi 3 Mini 128K Instruct has the following published scores: MMLU 68.7. Compare against reference models on this page or on the LLM Explorer leaderboards.

Phi 3 Mini 128K Instruct by microsoft — VRAM 7.7GB, 128K context

Name: Phi 3 Mini 128K Instruct
Rating: 3.91 (11 reviews)
Author: microsoft

Phi 3 Mini 128K Instruct is an open-source language model by microsoft. Features: 3.8b LLM, VRAM: 7.7GB, Context: 128K, License: mit, Instruction-Based, LLM Explorer Score: 0.36, Arc: 63.1, HellaSwag: 80.1, MMLU: 68.7, GSM8K: 69.5.

Code Conversational Custom code En Endpoints compatible Instruct Phi3 Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/microsoft/Phi-3-mini-128k-instruct

Phi 3 Mini 128K Instruct Benchmarks

ARC: 63.14 vs 96.7 (so35)^-34.7%

HellaSwag: 80.09 vs 95.3 (gpt4)^-16%

MMLU: 68.7 vs 88.3 (so35)^-22.2%

GSM8K: 69.52 vs 96.4 (so35)^-27.9%

LLME Score: 0.36229

^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)

Phi 3 Mini 128K Instruct (microsoft/Phi-3-mini-128k-instruct)

🌟 Advertise your project 🚀

Phi 3 Mini 128K Instruct Parameters and Internals

Model Type

text generation

Use Cases

Areas:

Research, Commercial applications

Primary Use Cases:

Memory/compute constrained environments, Latency bound scenarios, Strong reasoning tasks

Limitations:

Not specifically designed or evaluated for all downstream purposes.

Considerations:

Adherence to laws and regulations is required.

Additional Notes

This is a static model trained on an offline dataset with a cutoff date of October 2023. Future versions may improve upon it.

Supported Languages

English (Proficient)

Training Details

Data Sources:

Publicly available documents, Newly created synthetic data, High quality chat format supervised data

Data Volume:

4.9T tokens

Methodology:

Supervised fine-tuning, Direct Preference Optimization

Context Length:

128000

Training Time:

10 days

Hardware Used:

512 H100-80G GPUs

Model Architecture:

Dense decoder-only Transformer

Responsible Ai Considerations

Fairness:

Models can over- or under-represent groups, erase representation of some groups, or reinforce stereotypes.

Transparency:

Inappropriate or offensive content generation potential.

Accountability:

Developers need to ensure the model complies with laws and regulations.

Mitigation Strategies:

Use safety classifiers or implement custom safety solutions.

Input Output

Input Format:

Text in chat format

Accepted Modalities:

Text

Output Format:

Generated text in response to input

Performance Tips:

For certain GPUs, call AutoModelForCausalLM.from_pretrained() with attn_implementation="eager".

Release Notes

Version:

June 2024

Notes:

Improvement in long-context understanding, instruction following, reasoning capability.

LLM Name	Phi 3 Mini 128K Instruct
Repository 🤗	https://huggingface.co/microsoft/Phi-3-mini-128k-instruct
Model Size	3.8b
Required VRAM	7.7 GB
Updated	2026-05-20
Maintainer	microsoft
Model Type	phi3
Instruction-Based	Yes
Model Files	5.0 GB: 1-of-2 2.7 GB: 2-of-2
Supported Languages	en
Model Architecture	Phi3ForCausalLM
License	mit
Context Length	131072
Model Max Length	131072
Transformers Version	4.40.2
Tokenizer Class	LlamaTokenizer
Padding Token	<\|endoftext\|>
Vocabulary Size	32064
Torch Data Type	bfloat16

Quantized Models of the Phi 3 Mini 128K Instruct

Model	Likes	Downloads	VRAM
Phi 3 Mini 128K Instruct Gguf	1	1346	1 GB
Phi 3 Mini 128K Instruct GGUF	2	336	1 GB
Phi 3 Mini 128K Instruct GGUF	1	165	1 GB
...128K Instruct HQQ 2bit Smashed	0	8	1 GB
... 3 Mini 128K Instruct Bnb 4bit	1	7	2 GB

Best Alternatives to Phi 3 Mini 128K Instruct

Best Alternatives	Context / RAM	Downloads	Likes
Phi 4 Mini Instruct	128K / 7.7 GB	637798	706
Phi 3 Medium 128K Instruct	128K / 28 GB	8300	388
Phi 4 Mini Instruct ONNX GQA	128K / GB	52	6
Phi 3.5 Mini Instruct Onnx Web	128K / GB	726	15
Phi 3.5 Mini Instruct Onnx	128K / GB	108	37
Phi 3 Mini 128K Instruct Onnx	128K / GB	102	193
...Medium 128K Instruct Onnx Cuda	128K / GB	85	24
...i 3 Mini 128K Instruct Ov Int4	128K / 2 GB	6	0
... Medium 128K Instruct Onnx Cpu	128K / GB	67	14
...um 128K Instruct Onnx Directml	128K / GB	58	6

Note: green Score (e.g. "73.2") means that the model is better than microsoft/Phi-3-mini-128k-instruct.

Rank the Phi 3 Mini 128K Instruct 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? 54413 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.

Check out Ag3ntum — our secure, self-hosted AI agent for server management.

Release v20260328a

Support LLM Explorer

Phi 3 Mini 128K Instruct by microsoft