What are the hardware requirements for Phi 3 Medium 4K Instruct?

Phi 3 Medium 4K Instruct requires approximately 28 GB of VRAM and supports a context window of 4K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Phi 3 Medium 4K Instruct and how large is it?

Phi 3 Medium 4K Instruct is developed by unsloth, a model with 14b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Phi 3 Medium 4K Instruct perform on standard benchmarks?

Phi 3 Medium 4K Instruct has the following published scores: MMLU 77.85, LMSYS ELO 1197. Compare against reference models on this page or on the LLM Explorer leaderboards.

Phi 3 Medium 4K Instruct by unsloth — VRAM 28GB, 4K context

Name: Phi 3 Medium 4K Instruct
Author: unsloth

Phi 3 Medium 4K Instruct is an open-source language model by unsloth. Features: 14b LLM, VRAM: 28GB, Context: 4K, License: mit, Instruction-Based, LLM Explorer Score: 0.25, ELO: 1197, Arc: 67.1, HellaSwag: 85.7, MMLU: 77.9, GSM8K: 80.2.

Conversational Deploy:azure En Endpoints compatible Instruct Mistral Phi Phi3 Region:us Safetensors Sharded Tensorflow Unsloth

Model Card on HF 🤗: https://huggingface.co/unsloth/Phi-3-medium-4k-instruct

Phi 3 Medium 4K Instruct Benchmarks

LMSys ELO 1197 vs 1272 (so35) -5.9%

^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 Medium 4K Instruct Parameters and Internals

Model Type

quantized 4bit model

Use Cases

Primary Use Cases:

text completion, conversational templates

LLM Name	Phi 3 Medium 4K Instruct
Repository 🤗	https://huggingface.co/unsloth/Phi-3-medium-4k-instruct
Model Size	14b
Required VRAM	28 GB
Updated	2026-08-07
Maintainer	unsloth
Model Type	mistral
Instruction-Based	Yes
Model Files	4.9 GB: 1-of-6 5.0 GB: 2-of-6 4.9 GB: 3-of-6 5.0 GB: 4-of-6 5.0 GB: 5-of-6 3.2 GB: 6-of-6
Supported Languages	en
Model Architecture	MistralForCausalLM
License	mit
Context Length	4096
Model Max Length	4096
Transformers Version	4.52.4
Tokenizer Class	LlamaTokenizer
Padding Token	<\|placeholder6\|>
Vocabulary Size	32064
Torch Data Type	bfloat16

Rank the Phi 3 Medium 4K Instruct capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

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

Best Alternatives to Phi 3 Medium 4K Instruct

Best Alternatives	Context / RAM	Downloads	Likes
...ral Nemo Instruct 14B Merge V1	1000K / 24.6 GB	19	0
...2.9.2 Phi 3 Medium Abliterated	4K / 28 GB	9259	19
...2.9.2 Phi 3 Medium Abliterated	4K / 28 GB	8285	24
Phi 3 Medium Llamaish	4K / 28 GB	17	0
Phi3 Translator Merged3	4K / 55 GB	6	0
Arsh V1	4K / 28 GB	1123	1
...ll Phi 3 Medium 4K Inst Philos	4K / 28 GB	7	0
...antha2.0 Phi3 Medium Ita 16bit	4K / 28 GB	2604	0
Phi 3 Unsloth Finetune 7	4K / 28 GB	7	0
Phi 3 Unsloth Finetune 6	4K / 28 GB	8	0

Note: green Score (e.g. "73.2") means that the model is better than unsloth/Phi-3-medium-4k-instruct.