OpenAI CEO Sam Altman Says ChatGPT’s Best Days Are Gone

Sam Altman, the CEO of OpenAI, has acknowledged that the organization is no longer at the forefront of the artificial intelligence sector, particularly with the emergence of Chinese competitor DeepSeek and its R1 reasoning model. Nonetheless, he affirmed that OpenAI will still push forward in the industry.

During a recent Ask Me Anything (AMA) session on Reddit, Altman revealed that OpenAI has been “on the wrong side of history” regarding its open-source development approach for AI models. He, along with other executives, opened up about the company’s current situation and future strategies in a candid discussion with tech enthusiasts.

Altman expressed, “I personally think we have been on the wrong side of history and need to develop a new open-source strategy. However, not everyone at OpenAI agrees with this perspective, and it’s not our top priority at the moment.”

In response to a question about how DeepSeek might impact OpenAI’s development of future models, Altman stated, “It’s a very good model. We will create better models but will have a smaller lead than we did in previous years.”

OpenAI’s Chief Product Officer, Kevin Weil, commented on the potential of making some of their older models open-source, although he did not provide specific details regarding which models or when this could take place. “We’ll definitely consider doing more of this,” he remarked.

The entrance of DeepSeek has prompted OpenAI to reflect on its next steps. Following the recent launch of its o3-mini model, the company is contemplating a more transparent reasoning model, allowing users to observe its “thought process.” This capability is already present in DeepSeek’s R1, contributing to its appeal.

Altman and Weil also discussed rumors surrounding a potential price hike for ChatGPT, the AI chatbot that relies on most of OpenAI’s models. The CEO expressed hopes of eventually reducing ChatGPT’s pricing. Currently, the most expensive subscription tier costs $200 per month, while the Plus tier is priced at $20 monthly.

During the AMA, OpenAI unveiled teasers for upcoming products, including the next iteration of the o3 reasoning model, expected to launch within several weeks to months. While there is no confirmed release date for the next major model, which is likely to be labeled GPT-5, plans for a significant update of the DALL-E 3 image generator—a tool that hasn’t received major changes since its introduction two years ago—are also underway.

The executives addressed their collaborations with the U.S. government, particularly a data center initiative named Stargate. When questioned about the potential for OpenAI’s models to be utilized in the creation of destructive weapons, like nuclear arms, Weil expressed trust in the researchers involved not to misuse the technology.

This AMA follows a tumultuous period for OpenAI, sparked by DeepSeek’s announcement of its R1 reasoning model, which is said to compete with OpenAI and Meta’s offerings at significantly lower operating costs. DeepSeek is also free for users and open-source for developers, greatly enhancing its competitive edge.

While some speculate that DeepSeek may have leveraged questionable methods, such as distillation, to harvest data from OpenAI for its models, analysts suggest that the damage is done. Even former President Trump commented on the situation, calling it a “wake-up call” for the U.S. AI sector. Microsoft, a key investor in OpenAI, has shifted support to DeepSeek via Azure cloud and GitHub, promoting a democratization of AI models for its Copilot+ PCs.