For the past two years, Amazon has lagged behind Google and Microsoft in the generative AI field. Google, with its Gemini family of models, and Microsoft, with exclusive access to OpenAI's models, have attracted AI developers to their cloud platforms. Now, Amazon is catching up with the new Amazon Nova family of state-of-the-art foundation models, offering industry-leading price performance.
Since May, Google and OpenAI have been engaged in an LLM price war. Both companies have been constantly reducing prices of their latest large language models to make them more attractive for developers. Here's a quick timeline of what has been going on between Google and OpenAI:
- In May, Google announced the new Gemini 1.5 Flash model, which was aggressively priced ($0.35 per million input tokens and $1.05 per million output tokens) compared to other frontier models.
- In June, OpenAI announced the new GPT-4o Mini model to compete directly against Gemini 1.5 Flash, undercutting its pricing at $0.15 per million input tokens and $0.6 per million output tokens.
- In August, Google reduced the price of the Gemini 1.5 Flash model by about 80%, effective August 12, 2024. The new cost is $0.075 per million input tokens and $0.3 per million output tokens, making Gemini 1.5 Flash nearly 50% cheaper than OpenAI's GPT-4o mini.
- In August, OpenAI reduced the price of GPT-4o to $2.50 per 1 million input tokens and $10.00 per 1 million output tokens. At that time, the Gemini 1.5 Pro model had a price tag of $3.50 per 1 million input tokens and $10.50 per 1 million output tokens.
- In September, Google reduced the price of the Gemini 1.5 Pro model significantly. It now costs $1.25 per 1 million input tokens and $2.5 per 1 million output tokens.
- In October, Google introduced Gemini 1.5 Flash 8B, its cheapest AI model that costs $0.0375 per 1 million input tokens and $0.15 per 1 million output tokens.
With the new Nova family of models, Amazon has also joined the LLM price war. Amazon claims that Amazon Nova Micro, Amazon Nova Lite, and Amazon Nova Pro models are at least 75 percent less expensive than competing models in their respective classes in Amazon Bedrock. Here's how Amazon has priced its new Nova models:
- Amazon Nova Pro, the most powerful model from Amazon, is priced at $0.8 per 1 million input tokens and $3.2 per 1 million output tokens. When compared to OpenAI's GPT-4o, Nova Pro is priced at one-third the cost.
- Amazon Nova Lite is priced at $0.06 per 1 million input tokens and $0.24 per 1 million output tokens.
- Amazon Nova Micro is priced at $0.035 per million input tokens and $0.14 per million output tokens. This pricing even undercuts Google's cheapest model, Gemini Flash 8B.
The LLM market is becoming increasingly competitive both in terms of price and performance, with major players like Amazon, Google, and OpenAI vying for dominance. While lower prices benefit users, it remains to be seen how smaller AI startups will navigate this increasingly challenging landscape.
1 Comment - Add comment