Meta's Muse Spark 'Watermelon' Update Brings Stronger Coding and Agentic AI for Enterprise

Meta is preparing a major upgrade to its Muse Spark AI model, codenamed "Watermelon," that Chief AI Officer Alexandr Wang says will bring significant improvements to coding performance and autonomous agent capabilities. Wang announced the update via X, positioning it as a direct move to close the competitive gap with other leading AI platforms.

A Business Insider report citing anonymous sources said Watermelon — which uses substantially more compute than its predecessor — has already matched OpenAI's GPT 5.5 model in internal benchmarks.

Key Takeaways

• Enterprise cost pressure: Analysts say a stronger Muse Spark model would increase competition, lower AI costs, and give enterprises an alternative to OpenAI and Anthropic. "If offered as an open-weight or low-cost model, it could make AI coding assistants more affordable while improving data control and reducing vendor lock-in," said Pareekh Jain of Pareekh Consulting.
• Platform ambitions beyond models: Forrester analyst Charlie Dai noted that Meta appears to be moving "beyond foundation models" toward becoming a platform for AI-native applications and agents — pointing to the Pocket app and reported Manus acquisition efforts as signals of that direction.
• New API and infrastructure play: The updated model will be available through Meta AI and a new API. Meta is also reportedly developing cloud infrastructure business lines to sell AI compute and model access directly.

Enterprise adoption faces hurdles, however. Dai cautioned that Meta must demonstrate "superior real-world coding quality, reliable agent execution, strong security and governance, and a vibrant developer ecosystem" — particularly outside North America, where geopolitical and regulatory considerations increasingly shape model choices.

🔗 Read the full article on Computerworld