Key facts
- Meta Platforms has launched its Muse Spark 1.1 AI model for developers.
- The model is designed for coding and AI agent tasks.
- Pricing is set at $1.25 per million input tokens and $4.25 per million output tokens.
- CEO Mark Zuckerberg stated the model offers 'very low pricing' and outperforms Google's Gemini in several categories.
- Meta is increasing its capital expenditure guidance for the year to $125-$145 billion.
Meta Platforms has launched Muse Spark 1.1, a new multimodal AI model for developers focused on agentic coding tasks. This release positions Meta as a competitor to established players like OpenAI and Anthropic in the AI coding space. CEO Mark Zuckerberg emphasized the model's 'very low price' and strong performance, particularly in agentic tasks, tool use, and computer use, stating it outperformed Google's Gemini in several categories.
The pricing for Muse Spark 1.1 is set at $1.25 per million input tokens and $4.25 per million output tokens. This competitive pricing aims to attract enterprises seeking AI solutions for complex processes, bug fixing, and large code migrations. Meta's significant investment in AI is further underscored by an increased capital expenditure guidance for the year, projected between $125 billion and $145 billion.
This launch follows other recent AI announcements from Meta, including a new AI image-generation model called Muse Image, and coincides with releases from competitors such as SpaceXAI's Grok and OpenAI's GPT-5.6.