Announcing Amazon Nova 2.5 Sonic with improved reasoning for voice agents

This article announces the general availability of Amazon Nova 2.5 Sonic, an advanced speech-to-speech model for building natural, real-time voice agents with improved reasoning and lower latency.

• Enhanced reasoning, instruction following, and tool-calling accuracy for voice applications
• Enables voice agents to maintain context and complete multi-step customer tasks like order lookups and returns
• Supports seven languages with expressive voices and 256K context window
• Includes asynchronous tool calling and mixed voice/text sessions in same conversation
• Integrates with Strands Bidi Agents for production voice agents with minimal code
• Available in US East (N. Virginia), US West (Oregon), Europe (Stockholm), and Asia Pacific (Tokyo)
• Priced same as Nova 2 Sonic

Nova 2.5 Sonic enables developers to build more capable voice agents for customer service, personal AI assistants, and real-time conversational applications.