Why it matters: Amazon has launched Nova, a new family of AI models that process text, images and video at significantly lower costs than competitors. According to AboutAmazon, this major expansion of Amazon’s AI capabilities aims to make advanced artificial intelligence more accessible and affordable for businesses while challenging OpenAI and Google’s dominance in the space.
The Big Picture: According to Techcrunch, the Nova family includes six specialized models ranging from lightweight text processors to advanced multimedia generators:
- Nova Micro for fast, cheap text processing
- Nova Lite for basic multimedia tasks
- Nova Pro for complex reasoning
- Nova Premier coming in early 2025
- Nova Canvas for image generation
- Nova Reel for video creation
Cost Innovation: Amazon claims Nova models are up to 75% cheaper than competitors while maintaining high performance:
- Nova Micro: $0.000035 per 1,000 input tokens
- Nova Lite: $0.00006 per 1,000 input tokens
- Nova Pro: $0.0008 per 1,000 input tokens
Technical Capabilities: The models offer significant processing power, with Nova Lite and Pro capable of handling:
- Context windows up to 300,000 tokens
- Analysis of multiple images
- Up to 30 minutes of video
- 15,000 lines of code
Looking Forward: Amazon plans to expand the Nova family in 2025 with speech-to-speech models and advanced multimodal capabilities. The company’s focus on cost-effectiveness and integration with existing AWS services could accelerate AI adoption across industries. That could mean we will get the best desktop computers powered by AI that much sooner.
