Updated

Amazon Nova 2.5 Sonic

Amazon announced general availability of Nova 2.5 Sonic, a real-time speech-to-speech model for voice agents. The release adds improved reasoning, instruction following, tool-calling accuracy, and lower latency, according to AWS.

By No Code MBA · Facts checked

Provider
Amazon
Announcement date
Status at last check
Generally available
Model family
Amazon Nova
Preceding model
Amazon Nova 2 Sonic
Capabilities
Audio, Text

What changed

Nova 2.5 Sonic is a real-time speech-to-speech model for voice agents that AWS says improves reasoning, instruction following and tool-calling accuracy over Nova 2 Sonic, with lower latency. It keeps the earlier features: speech understanding, expressive voices in seven languages, controllable turn-taking, mixed voice and text in one session, asynchronous tool calling and a 256K context window. AWS did not publish latency numbers or benchmark scores in the announcement.

Access and availability

Generally available in Amazon Bedrock in US East (N. Virginia), US West (Oregon), Europe (Stockholm) and Asia Pacific (Tokyo). It works with Strands Bidi Agents, which is also generally available.

Official access documentation

What you can build

Evaluate it with representative voice-agent conversations that require turn-taking, multi-step instructions, and tool use, then review accuracy and handoff behavior.

Limitations

Available in only four Bedrock regions at launch. AWS gave no latency figures or benchmark results, and the seven supported languages are not listed in the announcement.

Sources and release history