Meta Challenges OpenAI: Watermelon Model Competes with GPT-5.5

⚡

Key Takeaways

1Alexandr Wang from Meta claims that their AI model Watermelon now matches OpenAI's GPT-5.5.

2Meta is ramping up its investments in AI to catch up with rivals like Google and Anthropic.

3Meta plans to spend up to $145 billion on infrastructure this year.

💡Why it matters — Meta is aiming to establish itself against OpenAI in the AI race, influencing innovation and competitiveness in the sector.

Meta Claims Its AI Watermelon Rivals GPT-5.5

Alexandr Wang, head of Meta's AI division, recently announced that their new artificial intelligence model, dubbed Watermelon, is now comparable to OpenAI's flagship model, GPT-5.5. This statement was made during an internal meeting, highlighting Mark Zuckerberg's ongoing efforts to bolster investments in the AI sector.

Meta aims to surpass OpenAI and other competitors in the artificial intelligence field. However, so far, these efforts have yet to yield the expected results. During the meeting, Wang explained that Meta is making significant progress in the race for AI models. He specified that the Watermelon model, the successor to the Avocado model, is currently under training. This new model utilizes computing power far superior to that of Avocado, which is the internal code name for Muse Spark, the first model in the series launched by Meta in April.

Meta's Ambitions Against Competition

Meta's ambitions in artificial intelligence are clear: to close the gap with giants like OpenAI, Google, and Anthropic. Despite considerable investments in chips, data centers, and recruiting talent, Meta struggles to convince developers and clients that its models are at the forefront of the industry.

If Wang's assessment proves accurate, it could indicate that Meta's investments and Zuckerberg's aggressive recruitment strategy are beginning to pay off. However, the AI race continues to evolve rapidly, and Meta still needs to prove its ability to compete with its rivals.

OpenAI and the Race for Innovation

OpenAI's GPT-5.5 model, a powerful chatbot, was launched in April of this year. OpenAI then unveiled its most advanced model to date, GPT-5.6, at the end of last month. However, the latter is not yet publicly accessible due to regulatory demands from the U.S. government.

Meta has chosen not to comment on these developments, while OpenAI has not responded to requests for comment. In April, Meta launched the first model in its series, Muse Spark, which demonstrated good performance on benchmarks but failed to match or surpass OpenAI or other labs like Anthropic.

Meta's Strategy to Dominate AI

Mark Zuckerberg is actively pushing for Meta to take a leading position in the race for artificial intelligence. Last year, he appointed Alexandr Wang to lead these efforts, renaming the company's AI division to Meta Superintelligence Labs.

Wang oversees an elite team of AI researchers known as TBD, along with other AI initiatives, including a recent advancement in hardware. Meta has reportedly offered AI experts hundreds of millions of dollars to attract them, according to a Business Insider report.

Massive Investments in Infrastructure

This talent quest comes as Meta ramps up its spending on infrastructure. The company has informed its investors that it plans to spend between $125 billion and $145 billion this year on chips, data centers, and other infrastructure. This forecast is an increase from a previous estimate of $115 billion to $135 billion, due to rising component costs and additional expenses for data centers.