Meta debuts Muse Spark 1.2 and first coding agent as it ramps up competition with OpenAI, Anthropic
Meta ( META ) debuted its latest AI model, Muse Spark 1.2, and Muse Code, the first coding agent from the company's Meta Superintelligence Labs (MSL), on Wednesday as the social media giant pushes toโฆ
Meta ( META ) debuted its latest AI model, Muse Spark 1.2, and Muse Code, the first coding agent from the company's Meta Superintelligence Labs (MSL), on Wednesday as the social media giant pushes to better compete with frontier AI labs like OpenAI ( OPAI.PVT ) and Anthropic ( ANTH.PVT ).
Muse Spark 1.2 is a coding-centric update to Meta's Muse Spark, which the company rolled out last month. Muse Code is the coding harness that runs Spark 1.2.
Coding is one of the biggest use cases for AI agents, helping to cut down on development time. Anthropic's Claude Code coding agent helped turn the company into an AI powerhouse and drove OpenAI and others to jump into the market.
While it's not a bleeding-edge frontier model, Meta does compare Muse Spark 1.2 to high-powered models like Anthropic's Opus 5, GPT 5.6 Terra, Gemini 3.6 Flash, and Grok 4.5.
Meta is in the midst of rebuilding its AI efforts after its Llama models failed to impress developers last year. Since then, CEO Mark Zuckerberg has gone on a hiring spree, bringing in Scale AI CEO Alexandr Wang in a $14.3 billion acquihire (Wang now serves as Meta's chief AI officer).
The company has been making steady progress on new models: In addition to Muse Spark, Spark 1.1, and Spark 1.2, the company has also debuted its Muse Image image generation model and previewed its Muse Video model.
But developing those models, including standing up the data centers necessary to run them, is costing Meta billions.
During its most recent earnings call, the company raised the lower end of its capital expenditure projections from between $124 billion and $145 billion to between $134 billion and $145 billion.
Read Full Story at Yahoo Finance โ
