Muse Spark 1.1 Scores 69, Rivals GPT-5.5 in Coding AI

Meta Superintelligence Labs has launched its upgraded coding agent model, Muse Spark 1.1. This model scored 69 on the Artificial Analysis Coding Agent Index, placing it close to GPT-5.5 and outperforming Claude Opus 4.8. Here's the thing: Muse Spark 1.1 is designed for agent-based applications, with a 1 million token context window. This allows it to handle large amounts of code in memory and supports delegating tasks to sub-agents for multi-step development work. Early adopters include Replit, Cline, and Box. What's interesting is the pricing. Muse Spark 1.1 is available through Meta’s first paid developer API at $1.25 per million input tokens and $4.25 per million output tokens, with new users receiving $20 in free credits. This makes it substantially cheaper than competitors. The bottom line: Meta's move into a paid API signals a strategic shift to capture revenue from high-performance AI services, potentially impacting the broader AI coding market.