Meta has announced that it is making substantial strides in the artificial intelligence sector, with its latest AI model, codenamed Watermelon, reportedly nearing the capabilities of OpenAI's flagship GPT-5.5. Alexander Wang, head of Meta's AI division, shared this update, although specific test benchmarks supporting his claims remain undisclosed. Watermelon, which is currently in training, is expected to utilize considerably more computational resources than its predecessor, Avocado, which was launched earlier this year. Wang also teased an upcoming update to the Muse Spark model, enhancing its coding and AI management abilities to further narrow the gap with industry leaders.
Meta aims to catch up with rivals such as OpenAI, Google, and Anthropic, having invested heavily in talent, infrastructure, and data centers. Despite these efforts, the company is still working to convince developers and clients of its models' competitiveness. If Wang's assertions hold true, it could signal a turning point for Meta in the fast-paced AI race, especially as OpenAI released GPT-5.5 in April and announced a more powerful GPT-5.6 shortly thereafter.
CEO Mark Zuckerberg's aggressive push for AI advancement has seen Wang appointed to lead this initiative, and the AI department has been rebranded as Meta Superintelligence Labs. With a team of top industry experts and a significant budget increase for infrastructure, Meta is determined to secure a leading position in the AI landscape. The outcome of these developments could reshape the competitive dynamics in the market, potentially altering the balance of power among AI developers.
Informational material. 18+.
