Meta has introduced Muse Spark 1.1, its latest artificial intelligence model, as part of a broader strategy to compete with leading AI firms such as OpenAI and Anthropic. This model is now available through a public API, allowing developers to integrate it into their applications, although access is currently limited to Meta's own platforms.
The pricing structure is designed to be competitive, starting with $20 in free credits and charging $1.25 per million tokens for input and $4.25 for output. Alexandr Wang, Meta's AI chief, emphasized that Muse Spark 1.1 excels in coding tasks, which are crucial for developing AI agents capable of performing multiple functions autonomously.
This update follows the recent release of Muse Image, indicating Meta's commitment to enhancing its offerings for creators and advertisers. The company is under pressure from investors to demonstrate a return on its substantial investments in AI, especially as it lacks a cloud infrastructure business like its competitors.
Wang noted that while Meta is shifting towards proprietary models, it remains committed to open-source initiatives. Additionally, a more powerful model, code-named Watermelon, is in development, although no release date has been provided