OpenAI launches GPT-Live-1 voice model for developers

Investing.com -- OpenAI released GPT-Live-1 in its API on Thursday, providing developers with a voice model designed for building voice-enabled applications and business workflows. The model can listen and speak simultaneously and can delegate reasoning and actions to other models and tools.

The API release includes features that allow developers to customize voice experiences for specific users and workflows. Language learning platform Speak reported that GPT-Live-1 reduced interruptions by nearly 80% compared to previous turn-based systems in early evaluations, giving learners more time to think before receiving responses.

GPT-Live-1 handles both incoming and outgoing audio through a single model, avoiding the delays associated with traditional speech-to-text, language model, and text-to-speech architectures. The model can delegate reasoning and tool calls to backend text models like GPT-6 Astra or third-party models.

Developers can shape an agent's tone, pace, and conversational style through system prompts. The model manages background noise and silence without interrupting conversations and maintains context retention across extended interactions. GPT-Live-1 supports telephony deployment for phone-based voice agents.

Tony Stoyanov, co-founder and chief technology officer at a healthcare company, stated that GPT-Live-1 simplified their code base by 80% and removed 23,000 lines of code compared to their previous system.

In performance evaluations, GPT-Live-1 improved Full Duplex Bench performance by 30 percentage points over GPT-Realtime-2.1. When paired with GPT-6 Astra at medium reasoning effort, it ranked first on Tau3, which measures voice-agent intelligence on end-to-end tasks.

OpenAI expanded its voice options for GPT-Live-1, adding new voices across different accents, dialects, and languages. The company plans to continue expanding voice options and language availability.

GPT-Live-1 is available in the API at $0.05 per minute for the front-end voice layer. Developers can pair it with their chosen backend model and agent framework

You May Also Be Interested In

• Microsoft plans data center push to triple its computing power - Bloomberg
• ARC Group to acquire Malaysia's Firstborn Top Capital in $1.09B deal
• Pathfinder Ventures Announces Appointment of Mickey Perret as Chief Executive Officer