New Azure Open AI models bring fast, expressive, and real‑time AI experiences in Microsoft Foundry

2026-02-25 · Source: Microsoft Foundry Blog articles · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Cloud Computing & IT Infrastructure · Depth: Intermediate, short

Summary

Microsoft Foundry is rolling out three new Azure OpenAI models: GPT-5.3-Codex, GPT-Realtime-1.5, and GPT-Audio-1.5. GPT-5.3-Codex, priced at $1.75/1M input tokens and $14.00/1M output tokens, offers 25% faster execution and unifies advanced coding with broader reasoning for long-running engineering tasks, supporting multi-step migrations, agentic developer workflows, and automated code reviews. GPT-Realtime-1.5 and GPT-Audio-1.5 enhance real-time voice interactions, showing a +5% lift on Big Bench Audio reasoning, +10.23% in alphanumeric transcription, and +7% in instruction following. These models feature more natural-sounding speech, higher audio quality, improved instruction following, and function calling support, suitable for conversational voice agents and hands-free workflows. GPT-Realtime-1.5 text input is $4.00/1M tokens, audio input $32.00/1M tokens; GPT-Audio-1.5 text input is $2.50/1M tokens, audio input $32.00/1M tokens.

Key takeaway

For NLP Engineers and CTOs building complex AI applications, these new Azure OpenAI models offer significant advancements in handling long-running tasks and real-time voice interactions. You should evaluate GPT-5.3-Codex for multi-step developer workflows and GPT-Realtime-1.5/GPT-Audio-1.5 for voice-first experiences requiring high accuracy and low latency, leveraging Microsoft Foundry's integrated evaluation and deployment capabilities to accelerate your projects.

Key insights

New Azure OpenAI models prioritize continuity and reliability for complex, real-time AI applications and long-running engineering tasks.

Principles

• AI systems benefit from sustained context and adaptability.
• Reliability and low latency are critical for real-time voice AI.

Method

The models integrate advanced coding with reasoning (GPT-5.3-Codex) and enhance speech understanding with function calling (GPT-Realtime-1.5, GPT-Audio-1.5) to support multi-step, context-aware interactions.

In practice

• Use GPT-5.3-Codex for large-scale code refactoring.
• Deploy GPT-Realtime-1.5 for low-latency voice agents.
• Automate code reviews with GPT-5.3-Codex.

Topics

• GPT-5.3-Codex
• Real-time Voice AI
• AI-assisted Coding
• Conversational AI
• Microsoft Foundry

Best for: NLP Engineer, CTO, VP of Engineering/Data, AI Engineer, Machine Learning Engineer, Software Engineer

Related on AIssential

• OpenAI’s New Voice Models Can Reason, Translate, and Transcribe – All While You’re Still Talking — OpenAI's new voice models enable real-time reasoning, translation, and transcription for highly interactive AI agents.
• Public preview: Voice-native agents in Microsoft Foundry — Microsoft's new Foundry Agent Service and Voice Live API simplify building real-time, speech-to-speech AI agents on Azure.
• Introducing OpenAI's newest chat model in Microsoft Foundry — GPT-chat-latest improves factual accuracy, tool calling, and response efficiency for production AI deployments.
• Sign of the future: GPT-5.5 — AI capabilities are rapidly accelerating across models, applications, and tool integration.
• OpenAI's NEW Voice Agent Model - GPT-RealTime 2 is dope! — OpenAI's new real-time voice models offer low-latency, bidirectional voice communication, translation, and transcription.
• OpenAI has yet another new coding model and this time it's really fast — OpenAI's new Codex-Spark model prioritizes real-time coding speed over raw accuracy using specialized Cerebras hardware.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Microsoft Foundry Blog articles.