Skip to content

Model Library

MicrosoftReleased April 23, 2024

Phi-3 Mini

Microsoft · Released April 23, 2024

Intelligence score

5

At a glance

Intelligence
5
Speed
—
Latency
—
Input price
Not published
Output price
Not published
Context window
3.0Mtokens
Parameters
3.8billions

Summary

Comparison summary

Phi-3 Mini Instruct 3.8B scores 5 (estimated) on the Artificial Analysis Intelligence Index, placing it well above average among other open weight non-reasoning models of similar size (median: 3).

Specifications

Technical specifications

Model namePhi-3 Mini
ProviderMicrosoft
Release dateApril 23, 2024
Intelligence score5
Speed—
Latency (TTFT)—
Input priceNot published
Output priceNot published
Context window3,000,000 tokens
Parameters3.8B
ReasoningNo
Input modalitiestext
Output modalities—
Open weightsYes
LicenseMIT
API availability—
SourceModel Intelligence legacy profile: phi-3-mini

FAQ · 16

Frequently asked questions

Phi-3 Mini Instruct 3.8B was released on April 23, 2024.

Phi-3 Mini Instruct 3.8B was created by Microsoft.

Phi-3 Mini Instruct 3.8B scores 5 (estimated) on the Artificial Analysis Intelligence Index, placing it well above average among other open weight non-reasoning models of similar size (median: 3).

No, Phi-3 Mini Instruct 3.8B is not a reasoning model. It provides direct responses without extended chain-of-thought reasoning.

Phi-3 Mini Instruct 3.8B supports text only input.

Phi-3 Mini Instruct 3.8B supports text only output.

No, Phi-3 Mini Instruct 3.8B does not support image input. It can only process text.

No, Phi-3 Mini Instruct 3.8B is not multimodal. It only supports text only input.

Phi-3 Mini Instruct 3.8B has a context window of 4.1k tokens. This determines how much text and conversation history the model can process in a single request.

Yes, Phi-3 Mini Instruct 3.8B is open weights. The model weights are publicly available and can be downloaded for self-hosting.

Phi-3 Mini Instruct 3.8B has 3.8 billion parameters.

Phi-3 Mini Instruct 3.8B is released under the MIT license. This license allows commercial use.

Phi-3 Mini Instruct 3.8B achieves a score of 5 on the Artificial Analysis Intelligence Index. This composite benchmark evaluates models across reasoning, knowledge, mathematics, and coding.

Phi-3 Mini Instruct 3.8B has a knowledge cutoff of October 2023. The model's training data includes information up to this date.

Phi-3 Mini Instruct 3.8B is an open weights model that can be self-hosted.

Phi-3 Mini Instruct 3.8B is an open weights model that can be downloaded and self-hosted.