OpenAI CEO Sam Altman stated that the cost to use a given level of artificial intelligence (AI) falls about 10 times every 12 months, and lower prices lead to much more use
Sam Altman, chief executive at OpenAI, stated that the cost to use a given level of artificial intelligence (AI) falls about 10 times every 12 months, and lower prices lead to much more use. The chatGPT creator shared this remark in his latest blog, “Three Observations.”
“You can see this in the token cost from GPT-4 in early 2023 to GPT-4o in mid-2024, where the price per token dropped about 150x in that time period. Moore’s law changed the world at 2x every 18 months; this is unbelievably stronger,” Altman said.
Notably, since the launch of ChatGPT in 2022, the cost of using large language models (LLM) has steadily decreased. In 2023, GPT-4 was priced at $0.03 per 1,000 input tokens and $0.06 per 1,000 output tokens. Later that year, GPT-4 Turbo lowered costs even further. In May 2024, GPT-4o dropped API pricing to $5 per million input tokens and $15 per million output tokens.
By mid-2024, GPT-4o Mini reduced costs even more to $0.15 per million input tokens. Newer models like o3 and o3-mini continued this trend, offering better performance at lower costs.
With significantly decreasing costs over the years, OpenAI has experienced remarkable growth in its user base. Within five days of its release in 2022, the start-up’s pioneer model ChatGPT gained 1mn users. By January 2023, it had reached 100mn monthly active users, making it one of the fastest-growing applications in history. As of December 2024, ChatGPT has around 300mn weekly active users, with over 1bn messages processed daily.
Also, new AI startups like DeepSeek are offering similar AI tools at even lower prices than OpenAI. DeepSeek-V3 is priced at $0.14 per million input tokens and $0.28 per million output tokens, compared to OpenAI o1 Preview's $15.00 and $60.00, respectively.
Altman also shared two other observations in his blog. He stated that the intelligence of an AI model roughly equals the log of the resources used to train and run it. “These resources are chiefly training compute, data, and inference compute. It appears that you can spend arbitrary amounts of money and get continuous and predictable gains; the scaling laws that predict this are accurate over many orders of magnitude,” he added.
In his third observation, Altman wrote about the socioeconomic value of increasing the intelligence of AI. He believes that there is super-exponential value in linearly increasing AI intelligence. “A consequence of this is that we see no reason for exponentially increasing investment to stop in the near future,” he added.
Towards the end of his blog Altman stated that if these three observations of his come true, it will have a significant impact on society.