Anthropic's Claude 3.7 Sonnet, costing only 'a few tens of millions of dollars' to train, highlights declining expenses in state-of-the-art AI model training. This contrasts sharply with the expenses for models like OpenAI's GPT-4 and Google's Gemini Ultra, which exceeded $100 million and $200 million respectively. Despite current affordability, Anthropic CEO Dario Amodei anticipates future models will incur training expenses in the billions, particularly as the industry progresses towards more sophisticated, long-duration reasoning models that require enhanced computational resources.
"Claude 3.7 Sonnet, cost 'a few tens of millions of dollars' to train using less than 10^26 FLOPs of computing power, indicating AI model training costs are decreasing."
"Amodei expects future AI models to cost billions of dollars, reflecting the increasing complexity and computing resources required for next-generation models."
Collection
[
|
...
]