phi-4 is a state-of-the-art open model built upon a blend of synthetic datasets, data from filtered public domain websites, and acquired academic books and Q&A datasets.
phi-4 models are available in gguf formats.
To run the smallest phi-4, you need at least 2 GB of RAM. The largest one may require up to 8 GB.
phi-4 models are available in gguf formats.
Phi-4 is provided under the MIT license.
Small, capable, and trained for high quality and advanced reasoning.
phi-4 is a state-of-the-art open model built upon a blend of synthetic datasets, data from filtered public domain websites, and acquired academic books and Q&A datasets. The goal of this approach was to ensure that small capable models were trained with data focused on high quality and advanced reasoning.
phi-4 underwent a rigorous enhancement and alignment process, incorporating both supervised fine-tuning and direct preference optimization to ensure precise instruction adherence and robust safety measures.
Phi-4 is designed to accelerate research on language models, for use as a building block for generative AI powered features.
Technical report: https://arxiv.org/pdf/2412.08905. Phi-4 is provided under the MIT license.
It provides uses for general purpose AI systems and applications (primarily in English) which require:
To understand the capabilities, Microsoft compares phi-4 with a set of models over OpenAI’s SimpleEval benchmark. At the high-level overview of the model quality on representative benchmarks. For the table below, higher numbers indicate better performance:
* These scores are lower than those reported by Meta, perhaps because simple-evals has a strict formatting requirement that Llama models have particular trouble following. Microsoft uses the simple-evals framework because it is reproducible, but Meta reports 77 for MATH and 88 for HumanEval on Llama-3.3-70B.
Create an account, open a chat, and start generating with Phi-4.
Sign in with your Google account and claim your starter credits.
Start a conversation with Phi-4 from the chat experience.
Every generation draws from your credits; top up anytime from the pricing page.
Quick answers about Phi-4, its license, and getting started.
phi-4 is a state-of-the-art open model built upon a blend of synthetic datasets, data from filtered public domain websites, and acquired academic books and Q&A datasets.
Phi-4 is designed to accelerate research on language models, for use as a building block for generative AI powered features.
To run the smallest phi-4, you need at least 2 GB of RAM. The largest one may require up to 8 GB.
Phi-4 is provided under the MIT license.
Create an account to get starter credits and open a chat with Phi-4. When your credits run out, choose a pack on the pricing page.
Sign in and start a conversation with Phi-4. Starter credits are included.
Phi 4