Amazon Nova 2.5 Sonic goes GA on Bedrock
Amazon's Nova 2.5 Sonic speech-to-speech model reached general availability on Bedrock on October 5 with 256K context, expressive voices in seven languages, asynchronous tool calling, and mixed voice/text sessions, at Nova 2 Sonic pricing.
Why it matters
Async tool calling lets voice agents fire off tasks like order lookups without stalling the conversation, plus Strands Bidi Agents GA for building production voice agents in a few lines of code.
What you could build with it
A startup could build an after-hours phone support agent for small clinics or retailers that looks up orders and initiates returns mid-conversation using asynchronous tool calls. A bank could deploy a voice assistant that reads confirmation codes and flight numbers one character at a time, exploiting the model's improved instruction following.
Does it hold up?
Too early to judge. GA began October 5 in four regions with no published independent benchmarks, so the improved reasoning and tool-calling claims remain vendor statements for now.
Built with Amazon Nova 2.5 Sonic goes GA on Bedrock
- Amazon's Nova 2.5 Sonic Voice Model Hits Bedrock in Four RegionsAI Weekly · Launch writeup covering GA regions, 256K context, seven languages, and the pricing-parity-with-Nova-2-Sonic detail.
- Announcing Amazon Nova 2.5 Sonic with improved reasoning for voice agentsAWS News Feed · Community announcement detailing async tool calling, Strands Bidi Agents integration, and regional availability.
Learn more
First spotted on article: source.
More AI models
OpenAI publishes 722 AI-generated math manuscripts from an unreleased modelOpenAI released 722 mathematical manuscripts in 372 result families to GitHub under Apache-2.0, produced by an…models · JEV 0.74Mistral Large 4 (Le Chonk): 1.05T-param open-weight MoEMistral's new flagship: a 1.05T-parameter mixture-of-experts multimodal model with 49B active params and 1M-token…models · JEV 0.72ElevenLabs Eleven v4 + v4 Turbo: new emotive TTS architecture with a 100ms real-time variantElevenLabs released Eleven v4 and Eleven v4 Turbo on September 28: a new text-to-speech architecture with inline…models · JEV 0.72VoiceStudio: open-source, fully-local ElevenLabs alternative with voice cloning, dubbing and dictation in 646 languagesVoiceStudio is an open-source, fully-local voice AI studio: voice cloning from a 3-second sample, voice design, video…models · JEV 0.71Mistral Forge: enterprise platform for training and continuously improving proprietary modelsMistral AI launched Forge (Sep 28) — a full-lifecycle model training platform (pre-training, SFT, DPO/ODPO, RL…models · JEV 0.69DeepSeek V4-Flash official API: public beta with upgraded agent capabilities, Responses API and Codex supportDeepSeek's official V4-Flash is now a public beta API with massively upgraded agent capabilities - benchmark scores…models · JEV 0.69
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.