LLM EXPLORER 60,645 MODELS INDEXED

Phi 3 Medium 128K Instruct by microsoft

By microsoft · 45634 downloads
HF Repo ↗ Copy URL

Phi 3 Medium 128K Instruct is an open-source language model by microsoft. Features: 14b LLM, VRAM: 28GB, Context: 128K, License: mit, Instruction-Based, LLM Explorer Score: 0.29, Arc: 66.5, HellaSwag: 84.9, MMLU: 76.8, GSM8K: 80.5.

  Code   Conversational   Custom code   Endpoints compatible   Eval-results   Instruct   Multilingual   Phi3   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/microsoft/Phi-3-medium-128k-instruct 

Phi 3 Medium 128K Instruct Parameters and Internals

Model Type 
text-generation, code
Use Cases 
Areas:
Commercial, Research
Applications:
Language and multimodal models, Generative AI
Primary Use Cases:
Memory/compute constrained environments for general AI system, Latency bound scenarios requiring strong reasoning, Research acceleration
Limitations:
Not evaluated for all downstream purposes, High risk scenarios may require debiasing techniques
Considerations:
Evaluate and mitigate accuracy, safety, and fairness before use in specific downstream scenarios.
Supported Languages 
English (primary)
Training Details 
Data Sources:
Publicly available documents, Newly created synthetic text data, High quality chat format supervised data
Data Volume:
4.8 trillion tokens
Methodology:
Supervised fine-tuning (SFT) and Direct Preference Optimization (DPO)
Context Length:
128000
Training Time:
42 days
Hardware Used:
512 H100-80G GPUs
Model Architecture:
Dense decoder-only Transformer
Responsible Ai Considerations 
Fairness:
Trained primarily on English text, under-representation of some English language varieties is possible.
Transparency:
Models may generate inappropriate or offensive content; transparent communication with users recommended.
Accountability:
Developers should adhere to applicable laws and verify outputs before use in high-risk scenarios.
Mitigation Strategies:
Built-in safety classifiers and custom solutions for high-risk contexts.
Input Output 
Input Format:
Chat format prompts
Accepted Modalities:
Text
Output Format:
Generated text
Performance Tips:
Add a BOS token (`~~`) at the start of the conversation for reliable results.
Release Notes 
Version:
May 21, 2024
Date:
2024-05-21
Notes:
Phi-3-Medium-128K-Instruct model released with 14B parameters and advanced capabilities.
LLM NamePhi 3 Medium 128K Instruct
Repository 🤗https://huggingface.co/microsoft/Phi-3-medium-128k-instruct 
Model Size14b
Required VRAM28 GB
Updated2026-07-26
Maintainermicrosoft
Model Typephi3
Instruction-BasedYes
Model Files  4.9 GB: 1-of-6   5.0 GB: 2-of-6   4.9 GB: 3-of-6   4.8 GB: 4-of-6   4.8 GB: 5-of-6   3.6 GB: 6-of-6
Model ArchitecturePhi3ForCausalLM
Licensemit
Context Length131072
Model Max Length131072
Transformers Version4.39.3
Tokenizer ClassLlamaTokenizer
Padding Token<|endoftext|>
Vocabulary Size32064
Torch Data Typebfloat16

Quantized Models of the Phi 3 Medium 128K Instruct

Model
Likes
Downloads
VRAM
...hi 3 Medium 128K Instruct GGUF24945 GB
...hi 3 Medium 128K Instruct GGUF03175 GB

Best Alternatives to Phi 3 Medium 128K Instruct

Best Alternatives
Context / RAM
Downloads
Likes
...ess V2.5 Phi 3 Medium 128K 14B128K / 28 GB164
...i 3 Medium 128K Instruct C4 32128K / 15.5 GB50
Finetuned Phi3 Medium 128K128K / 28 GB80
Shisa V1 Phi3 14B128K / 28 GB52
Mahou 1.2 Phi 14B128K / 28 GB81
...colatine 14B Instruct DPO V1.316K / 29.4 GB160
Phi4 Slerp2 14B16K / 28 GB50
Phi 3 Medium 4K Instruct4K / 28 GB14252226
...colatine 14B Instruct DPO V1.24K / 28 GB2314
Neos Phi 3 14B V0.14K / 28 GB50
Note: green Score (e.g. "73.2") means that the model is better than microsoft/Phi-3-medium-128k-instruct.