VENDORiQ: OpenAI Launches New Features

Boosts from an already supercharged OpenAI GPT-4 Turbo

The Latest

In November 2023, OpenAI introduced the following new features:

  • GPT-4 Turbo includes knowledge up to April 2023, a 128k context window to accommodate over 300 pages of text
  • Function calling now allows multiple actions in a single message
  • Improved instruction following and JSON mode
  • New seed parameter in the beta feature for reproducible outputs and consistent model completions
  • New version of GPT-3.5 Turbo that supports a 16K context window by default
  • Assistants API designed to support developers through agent-like interactions

Why It’s Important

Enterprise leaders considering the deployment of large language models (LLMs) need to understand the significance of context windows in maintaining relevance and coherence in extended interactions. These are critical in ensuring stable and consistent responses to reduce the risk of erratic or off-topic outputs and in maintaining reliable AI-assisted communication channels.

OpenAI also announced cheaper pricing for input and output tokens of GPT-4 Turbo compared to GPT-4. This will encourage more experimentation and innovation since organisations can better afford to explore new ways to leverage AI. More enterprises can also deploy AI solutions extensively and frequently while maintaining cost efficiency.

Who’s Impacted

  • CTO
  • AI developers
  • IT teams

What’s Next?

Enterprises should be aware that while this new version has significantly enhanced capabilities, it may also amplify existing drawbacks associated with LLMs. Enterprises should still conduct a thorough evaluation of the potential risks and drawbacks associated with its use. This includes issues related to accuracy, ethical concerns, or unintended biases in the AI’s responses.

Related IBRS Advisory

  1. The Top Six Risks of Generative AI
  2. Generative AI for Enterprise Use – An Overview of the State-of-the-Art

Trouble viewing this article?

Search