Meta publicly launched a new generation of Muse Spark 1.1, further competing for the AI programming model. The model is oriented towards a proxy coding landscape that supports multi-step reasoning, tool call-up and cross-application task processing, with targeted clients including users requiring automated development in synergy with business processes.

Main-play proxy code and business process

Meta states that Muse Spark 1.1 can handle more complex digital workflows and deploy new features in enterprise systems. According to the company, this model is prominent in individual agency tasks requiring the planning, organization and mobilization of external applications and services.

In product positioning, Muse Spark 1.1 is not just a code completion tool, but a programming agent closer to a multi-step task. It aims at the direction closer to the AI programming assistant that OpenAI and Anthropic have been pushing for some time.

Pricing for the same product

Reuter quoted pricing information that Meta charged Muse Spark 1.1:

  • Token per million
  • Token per million
  • Pricing is close to the same models as Anthropic, OpenAI

This pricing, by reported calibre, is largely in the same competition range, slightly higher than Claude Haiku 4.5 in Anthropic and close to the GPT-5.6 Luna in OpenAI. For the current AI model market, prices remain one of the important factors in the choice of products by business clients.

Meta accelerates the completion of the product Line

Muse Spark's first version was released in April this year, and this is a new version on the public line. Although Meta started later than some of its counterparts in the AI programming model area, it continued to expand the base model product line, indicating that the company was accelerating the completion of related capabilities.

Meta CEO Zuckerberg is also writing on social platform X to build momentum for the model. It's about three years later he posted on the platform again. He described Spark as a low-priced model that performed well in terms of agency capacity, tool use and computer operations, and referred to “more content soon”, suggesting that Meta might continue to introduce new models.

The AI Model release continues to heat up.

This week, the AI industry is very busy with new products. Meta also launched the image generation model Muse Image on Tuesday. At the same time, OpenAI has published the GPT-5.6 series, while XAI has a new version of Grok.

In the context of the gradual convergence of modelling capacity, product pricing, task execution capacity and the speed of business landing are becoming important directions for producers in their competition for markets. Meta's introduction of Muse Spark 1.1 also means that competition on the AI programming track has increased.