Meta launched a new version of Muse Spark, a multimodal AI model aimed at agentic coding, on Thursday. This model is intended to compete with offerings from OpenAI and Anthropic.
Muse Spark 1.1, which follows the initial announcement in April, can perform multistep reasoning, manage complex processes, and govern digital workflows within enterprise systems, according to the company. Despite being perceived as lagging behind its competitors, Meta’s entry into this segment poses a potential challenge to established players.
The pricing of Muse Spark reflects an ongoing competitive strategy within the AI sector. Reuters reported that Meta will charge $1.25 per million input tokens and $4.25 per million output tokens, positioning its rates slightly above those of Anthropic’s Claude Haiku 4.5 and OpenAI’s GPT-5.6 Luna.
Meta promotes Spark as capable of managing large agentic workloads, resolving bugs, and assisting with substantial code migrations. The company stated, “Muse Spark 1.1 delivers exceptional performance in personal agentic tasks that require planning and orchestration across a range of external apps and services.”
CEO Mark Zuckerberg made a noteworthy return to X, posting for the first time in three years to discuss the launch. He characterized Spark as “a strong agentic and coding model at a very low price,” highlighting its strengths in agentic performance, tool use, and computer use. He also hinted at future developments, stating there would be “more to come soon.”
This week has seen significant activity in AI announcements for Meta, which also introduced a new AI image-generation model called Muse Image on Tuesday. In addition, Thursday marked the release of a new Grok version from SpaceXAI and a new family of models, GPT-5.6, from OpenAI. The competitive dynamics within the AI industry remain intense, with companies vying to differentiate their offerings.
