Announcing Amazon Nova 2.5 Sonic with improved reasoning for voice agents
Amazon's own voice AI model just got smarter and faster at handling real multi-step phone/voice tasks, like looking up an order or processing a return, without losing the plot partway through. If your business runs customer-service calls or voice bots on AWS, this is the model your tech team would reach for.
What changed
Source summary
AWS announced general availability of Nova 2.5 Sonic, a speech-to-speech model for real-time voice agents on Amazon Bedrock. It improves reasoning, instruction-following, and tool-calling accuracy and reduces latency versus Nova 2 Sonic, enabling multi-step tasks like customer-service lookups and returns. It keeps Nova 2 Sonic features (multilingual voices, turn-taking control, voice and text in one session, async tool calls, 256K context window) and is available in four Bedrock regions (US East N. Virginia, US West Oregon, Europe Stockholm, Asia Pacific Tokyo) at the same pricing as Nova 2 Sonic.
Why it matters to your work
Practical interpretation
Amazon's own voice AI model just got smarter and faster at handling real multi-step phone/voice tasks, like looking up an order or processing a return, without losing the plot partway through. If your business runs customer-service calls or voice bots on AWS, this is the model your tech team would reach for.
What to try next
If your business runs customer-service voice bots on AWS, ask your tech team to trial Nova 2.5 Sonic for faster, more accurate handling of multi-step calls like order lookups and returns.
Source & availability
This is a summary of a published provider announcement, with practical context from SCALA’s feed.
Plan, region, workspace and rollout conditions may affect access in Australia. Check the original source for current availability.
aws.amazon.com