Meta's Muse Spark 1.3 Launches: Frontier Performance, Competitive Pricing, and a Top-3 Ranking
Key Info
Meta has launched Muse Spark 1.3, now available in Muse Code and the Meta Model API. The company calls it its most capable model yet, with frontier-level performance, stronger agentic and coding abilities, and unusually low pricing; early community benchmarks place it at #3 overall, the first model to sit between Claude and GPT, with top results on DeepSWE.
Highlights
- Muse Spark 1.3 is available today in Muse Code and the Meta Model API, with max reasoning to follow after safety testing.
- Meta highlights major gains in agentic tasks, coding, and usability.
- The model reportedly ranks #3, between Claude and GPT, and leads DeepSWE at 75.4% ahead of GPT-5.6 Sol and Fable 5.
- Its pricing compared with Fable 5 is described as a "joke" — a surprisingly strong value proposition.