Meta challenges OpenAI with its first paid AI API, debuting Muse Spark 1.1
Meta opened its AI model Muse Spark to developers and released an upgraded version called Muse Spark 1.1 on Thursday, according to Meta. This is the first time that external developers can use the model through Meta’s public API. With this launch, Meta competes directly with artificial intelligence companies OpenAI and Anthropic By charging developers to use its artificial intelligence models.
Meta describes Muse Spark 1.1 as the strongest AI model for coding and agent tasks. Agent AI means models can complete tasks on their own with less human help. Meta says Muse Spark 1.1 can write new code, fix coding errors, understand software and solve programming problems. The model can understand text, images, and videos simultaneously because it is a multi-modal AI model. The launch signals Meta’s efforts to compete with OpenAI and Anthropic in the rapidly growing artificial intelligence market.
Muse Spark 1.1 Features
AI can also automate tasks using software and external tools, Meta said. The company says Muse Spark 1.1 can perform difficult multi-step tasks without the need for ongoing human instructions. Meta launched its first Muse Spark inference model in April after creating a superintelligence team last year to compete with leading AI companies.
Metamodel API launch
The company also launched a public preview of the Meta Model API, allowing developers to build applications using Muse Spark. According to Reuters, the API works like a digital bridge, allowing developers to connect Meta’s artificial intelligence models to their own software. Developers in the United States can now test Muse Spark through the metamodel API.
Also read: OpenAI launches GPT-Live speech model that can speak and listen at the same time
They can test prompts, compare AI responses and build new applications before launching them. Every new developer who signs up receives $20 in free credits to test the model before paying for it. After the free credits end, Meta charges $1.25 for every 1 million input tokens and $4.25 for every 1 million output tokens. This means that AI can remember and process longer documents and conversations at once. According to Reuters, Meta is priced higher than OpenAI’s GPT-5 mini and Anthropic’s Claude Haiku 4.5, but cheaper than Anthropic’s Claude Sonnet 4.6.
Artificial Intelligence Applications and Coding
Muse Spark 1.1 now available in Think mode meta-artificial intelligence Application and Meta’s website. Meta also plans to replace the Llama AI model currently used in WhatsApp, Instagram, Facebook and Meta smart glasses with Muse Spark. Earlier this week, Meta also launched Muse Image, Meta Superintelligence Labs’ first image generation model.
Meta says Muse Spark 1.1 comes with a massive one million token context window, allowing it to understand and remember more information in a single conversation. The company says the model is designed to work in multi-agent systems, where multiple AI agents can work together to complete the same task. Meta said that Muse Spark 1.1 can act as both a primary AI agent and a secondary AI agent. This allows the model to divide large tasks into smaller jobs and complete them faster by running them in parallel.
Better encoding performance
Meta says the upgraded model performs much better than earlier versions on large enterprise coding projects. Meta says AI can automatically find software bugs, fix them, add new features to old software, and help companies migrate large amounts of code to new systems. Meta also introduces new computer usage features in Muse Spark 1.1.
Artificial intelligence can automate work across different desktop applications by creating your own automation scripts. If scripting is not possible, AI can interact directly with software interfaces to complete tasks. Some companies have started using the API since its inception.
Security and Early Partners
The first major partners include Replit, Cline and Box. According to Investing.com, Replit CEO Amjad Masad said the combination of a million-coin context window, multi-mode capabilities and advanced coding capabilities could change the way developers build software at scale.
Cline CEO Saoud Rizwan said the model’s flexible tooling and competitive pricing make it attractive to developers handling large coding workloads. Yashodha Bhavnani, Vice President of Artificial Intelligence Products at Box said Spark of Muse 1.1 According to Investing.com, in Box’s internal testing, its performance was on the same level as some of the industry’s leading AI models.
Meta said it also conducted extensive security testing before releasing the model. Security checks look for risks involving chemical and biological misuse, cybersecurity threats, and situations in which artificial intelligence systems could lose control. Meta says Muse Spark 1.1 remains within safe limits and a detailed security report will be released soon.
The launch of Muse Spark 1.1 and Muse Image shows that Meta is expanding its artificial intelligence products for consumers and enterprises, while intensifying competition with OpenAI and Anthropic.