Meta Launches Its First Advanced AI Model "Muse Wonder"

Meta announced the launch of its first AI model in the "Frontier" class under the name "Muse Wonder," as reported by TechCrunch. The model was developed by Meta's Superintelligence Lab, which also launched an expanded version named "Muse Wonder Full."

Muse Wonder is available through an API called the "Meta Model API," priced at $4 per million input tokens and $20 per million output tokens. Muse Wonder Full, however, will only be available to Muse Code subscribers and will not be served through the API. The new release ranks higher than the previous "Muse Spark" model, positioning it at the forefront of the company's offerings.

Meta stated that both models will power "Muse," an AI agent being developed by the company. This means the new model will not be limited to answering questions, but will also manage multiple tasks across the company's products.

The launch comes amid an accelerating race among major tech companies to build more powerful and cheaper-to-run models. Token pricing determines the cost of using the model, making low pricing an attractive factor for developers and businesses.

What do these terms mean?

Frontier Model: An AI model at the highest levels of capability and performance, developed using state-of-the-art technologies, and typically the most expensive to develop and operate.

Token: The smallest unit of text processed by a model, which may be a word or part of a word; usage pricing is calculated based on their count.

API: A software intermediary that allows developers to use a service or model without building its technology from scratch.