Microsoft: Phi-3 Mini 128K Instruct

Text input Text output
Author's Description

Phi-3 Mini is a powerful 3.8B parameter model designed for advanced language understanding, reasoning, and instruction following. Optimized through supervised fine-tuning and preference adjustments, it excels in tasks involving common sense, mathematics, logical reasoning, and code processing. At time of release, Phi-3 Medium demonstrated state-of-the-art performance among lightweight models. This model is static, trained on an offline dataset with an October 2023 cutoff date.

Key Specifications
Cost
$$
Context
128K
Parameters
3.8B (Rumoured)
Released
May 25, 2024
Speed
Ability
Reliability
Supported Parameters

This model supports the following parameters:

Tool Choice Top P Temperature Tools Max Tokens
Features

This model supports the following features:

Tools
Performance Summary

Microsoft's Phi-3 Mini 128K Instruct demonstrates exceptional speed, consistently ranking among the fastest models available. It also offers highly competitive pricing, making it a cost-effective solution. Furthermore, the model exhibits outstanding reliability, providing usable responses with minimal technical failures, placing it in the 98th percentile. In terms of performance across categories, Phi-3 Mini shows a mixed profile. It excels in the Ethics (98.0% accuracy) and General Knowledge (93.4% accuracy) benchmarks, indicating strong capabilities in these areas. Its Email Classification accuracy is also respectable at 87.0%. However, the model struggles significantly with Instruction Following, achieving 0.0% accuracy, which is a critical weakness. Performance in Coding (59.0% accuracy) and Reasoning (38.0% accuracy) is below average, suggesting limitations in complex problem-solving and programming tasks. While its speed and cost efficiency are significant advantages, the model's inconsistent accuracy across different domains, particularly in instruction following and complex reasoning, should be considered for specific application requirements.

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.1
Completion $0.1

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
Azure
Azure | microsoft/phi-3-mini-128k-instruct 128K $0.1 / 1M tokens $0.1 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Free Executions Accuracy Cost Duration
Other Models by microsoft