amazon-nova

AWS has developed Reverse Direct Preference Optimization (rDPO), a novel unlearning technique that reduces over-deflection in Amazon Nova models by up to 53 percentage points. The approach allows organizations to selectively adjust content moderation safeguards while preserving general model capabilities through LoRA adapters.

AWS released guidance on fine-tuning Amazon Nova Micro and Nova Lite models for automated email data extraction using SageMaker AI. In collaboration with Parcel Perform, the fine-tuned Nova Micro achieved 94.77% extraction accuracy—a 16.6 percentage point improvement—while reducing inference costs by 50% and latency by 30% compared to previous models.

AWS published a technical demonstration showing that pairing Amazon Nova 2 Lite with Anthropic's Claude Sonnet 4.6 reduces document processing costs by approximately two-thirds compared to single-model approaches. The two-stage pipeline processed 336 scanned yearbook pages at $0.0027 per page, producing 3,122 name-to-face associations with 93% scoring at or above 0.95 confidence.

AWS has published a technical guide on hyperparameter optimization for Amazon Nova Forge, its platform for building custom frontier models from Amazon Nova checkpoints. The guide addresses three core challenges: catastrophic forgetting during domain specialization, learning rate calibration when mixing proprietary and curated training data, and baseline performance constraints for reinforcement fine-tuning.

Popsa, a photo book service operating in 50+ countries, generated over 5.5 million AI-powered titles in 2025 using Amazon Nova models. The company achieved 73% positive user feedback with Nova Pro while reducing costs and latency compared to Claude 3 Haiku.

Amazon Bedrock now supports Reinforcement Fine-Tuning (RFT) for customizing Amazon Nova and open source models using reward signals instead of labeled datasets. AWS reports up to 66% accuracy improvements over base models with reduced customization complexity. The approach works best for tasks with verifiable correctness (code, math) or subjective evaluation (moderation, summarization).

Amazon has published a technical guide for building real-time conversational podcasts using Amazon Nova 2 Sonic, its speech understanding and generation model. The solution demonstrates streaming audio generation, multi-turn dialogue between AI hosts, and stage-aware content filtering through a web interface.

Amazon Bedrock has introduced three distinct video analysis workflows that leverage multimodal foundation models to extract insights from video content at scale. The approaches—frame-based, shot-based, and multimodal embedding—are designed for different use cases and cost-performance trade-offs, with open-source reference implementations available on GitHub.

Amazon Bedrock now enables reinforcement fine-tuning (RFT) across multiple model families including Amazon Nova, open-weight models like OpenAI's GPT-OSS 20B, and Qwen 3 32B. The service automates the end-to-end customization workflow using GRPO optimization, allowing models to learn from feedback on multiple responses rather than static training datasets, with support for OpenAI-compatible APIs.

Amazon Nova 2 Lite expands context window to 1 million tokens, introduces extended thinking with developer controls, and adds native tool use and web grounding. AWS claims Nova 2 Lite surpasses Nova 1 Pro on multi-step reasoning while costing 7x less and running up to 5x faster.