×
Microsoft Launches Serverless Fine-Tuning for Its Phi-3 Model
Written by
Published on
Join our daily newsletter for breaking news, product launches and deals, research breakdowns, and other industry-leading AI coverage
Join Now

Microsoft introduces serverless fine-tuning for its Phi-3 small language model, allowing developers to adapt the AI for specific use cases without managing their own servers, intensifying competition in the enterprise AI market.

Key features of Phi-3 small language model: Microsoft’s Phi-3, a 3 billion parameter model, offers affordable performance on coding, common sense reasoning, and general knowledge, now as part of a family of 6 models with varying parameters and context lengths:

  • Phi-3 models range from 4,000 to 128,000 tokens in a single input, with costs from $0.0003 to $0.0005 USD per 1,000 input tokens.
  • Designed with safety features for enterprise use, Phi-3 models can be fine-tuned for specific verticals and use cases.

Serverless fine-tuning expands capabilities: The introduction of serverless fine-tuning for Phi-3-mini and Phi-3-medium models allows developers to create customized versions without managing infrastructure:

  • Developers can fine-tune models with their own data to build more relevant AI experiences safely and economically.
  • Serverless fine-tuning is well-suited for scenarios like learning new skills, enhancing response consistency and quality, and improving base model performance.
  • Educational software company Khan Academy is already using a fine-tuned Phi-3 to benchmark performance for its Khanmigo for Teachers application.

Intensifying competition in the enterprise AI market: Microsoft’s move comes amidst a flurry of activity from competitors like OpenAI, Meta, and Mistral, highlighting the race to capture enterprise AI developers:

Analyzing the implications: Microsoft’s introduction of serverless fine-tuning for Phi-3 marks a significant step in making AI more accessible and customizable for enterprises, but it also raises questions about the long-term sustainability of free and low-cost offerings as competition heats up. As the AI arms race continues, it remains to be seen how these developments will impact the broader landscape of enterprise AI adoption and innovation.

Microsoft unveils serverless fine-tuning for its Phi-3 small language model

Recent News

MILS AI model sees and hears without training, GitHub code released

Meta researchers develop system enabling language models to process images and audio without specialized training, leveraging existing capabilities through an innovative inference method.

Mayo Clinic combats AI hallucinations with “reverse RAG” technique

Mayo's innovative verification system traces each AI-generated medical fact back to its source, dramatically reducing hallucinations in clinical applications while maintaining healthcare's rigorous accuracy standards.

Columbia dropouts launch Cluely, an AI tool designed for cheating in interviews and exams

Columbia dropouts' desktop AI assistant provides real-time answers during interviews and exams through an overlay invisible during screen sharing.