Phi-2

Name: Phi-2
Author: Microsoft Azure

Foundation model

Phi-2

Release date: 2023-12-12Updated: 2024-01-06 19:45:14.535822

Live demoGitHubHugging Face Compare

Parameters

2.7B

Context length

Chinese support

Not supported

Reasoning ability

Phi-2 is an AI model published by Microsoft Azure, released on 2023-12-12, for Foundation model, with 2.7B parameters, and 2K context length, requiring about 5.4GB storage, with a 61.10 score on GSM8K.

Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

Phi-2

Model basics

Reasoning traces

Not supported

Thinking modes

Thinking modes not supported

Context length

2K tokens

Max output length

No data

Model type

Foundation model

Modality (in / out)

No data

Release date

2023-12-12

Model file size

5.4GB

MoE architecture

Total params / Active params

2.7B / N/A

Knowledge cutoff

No data

Phi-2

Open source & experience

Code license

MIT License

Weights license

MIT License- Commercial use permitted

GitHub repo

GitHub link unavailable

Hugging Face

https://huggingface.co/microsoft/phi-2

Live demo

No live demo

Phi-2

Official resources

Paper

DataLearnerAI blog

No blog post yet

Phi-2

API details

API speed

No data

No public API pricing yet.

Phi-2

Benchmark Results

Phi-2 currently shows benchmark results led by MBPP (29 / 70, score 59.10), GSM8K (34 / 70, score 61.10), HumanEval (53 / 101, score 48.30). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.