Meta’s Superintelligence Labs chief Alexandr Wang revealed in an internal all-hands meeting that the company’s next-generation large language model — codenamed “Watermelon” — has matched OpenAI’s flagship GPT-5.5 on benchmark tests.
The model is the successor to Muse Spark (internal codename “Avocado”), which Meta released in April. Wang said the new model uses an order of magnitude more compute than its predecessor.
Wang also posted on X that Muse Spark is getting an update soon, with major improvements in coding and agent capabilities. When a user asked when it would catch up to Anthropic’s Claude Opus, Wang replied: “Soon.”