Mark Zuckerberg reappears on X to announce Muse Spark 1.1 and sparks an AI price war against OpenAI and Anthropic

In Short

Meta is challenging Anthropic and OpenAI with its first AI-powered coding tool: Muse Spark 1.1. The launch is significant enough that Meta CEO Mark Zuckerberg broke a three-year silence on X to announce it. With this move, Zuckerberg has also ignited a price war in the AI coding tool sector.

Mark Zuckerberg reappears on X to announce Muse Spark 1.1 and sparks an AI price war against OpenAI and Anthropic
X
Font size
FOLLOW ON Google News

Meta is looking to regain ground in the AI race and is betting big on its latest tool: Muse Spark 1.1. Meta CEO Mark Zuckerberg announced the launch of Muse Spark 1.1 in his first post on X since July 2023—a three-year hiatus—underscoring the high stakes involved. However, the AI coding landscape already features established competitors like Claude Code and OpenAI's Codex. To lure developers to Muse Spark 1.1, Zuckerberg appears to have an ace up his sleeve: pricing.

In announcing Muse Spark 1.1 on X, Mark Zuckerberg described the new tool as "a powerful model capable of acting autonomously [agentic] and coding, at a very low price."

These statements come at a time when rising costs for AI models—whether from Anthropic or OpenAI—are prompting companies to rethink their usage. Some companies, such as Uber and Walmart, have even gone so far as to limit the number of AI tokens (the basic units of measurement for these models) their employees can use.

But before diving into the pricing of Muse Spark 1.1, let’s examine the capabilities of this new coding tool from Meta.

Meta claims Muse Spark 1.1 outperforms competitor models

Meta asserts that Muse Spark 1.1 is one of the most powerful coding models on the market. It is designed to handle workloads that companies increasingly delegate to AI systems, such as error correction, digital workflow management, implementing new features in corporate systems, and large-scale code migrations.

Muse Spark 1.1 was developed by Meta's Superintelligence Labs (MSL), led by Alexandr Wang.

According to Wang, the model was trained to perform well in programming, thereby enhancing the capabilities of AI agents to autonomously perform multiple tasks. "In a way, you have to develop programming capabilities as part of that, based on the agent's general capabilities," he told CNBC. He added that the model was trained to work with popular tools already used by developers, describing this as the best approach to maximise adoption.

When Mark Zuckerberg announced Muse Spark 1.1 on X, he shared performance comparisons between various AI models regarding programming tasks. According to the results, Muse Spark 1.1 outperforms Anthropic's Opus 4.8 and OpenAI's GPT-5.5 in certain tasks. It is worth noting that these comparisons did not include Anthropic's Mythos and Fable models or OpenAI's GPT-5.6.

Muse Spark 1.1 is more cost-effective than its rivals; Elon Musk's reaction

A key aspect Meta highlights regarding Muse Spark 1.1 is its pricing. Developers can access Muse Spark 1.1 via the Meta Model API. According to the company, each new API account starts with $20 in free credits. Subsequently, the cost is $1.25 per million input tokens and $4.25 per million output tokens.

To put this in perspective, Grok 4.5—the new AI model launched just a day ago by SpaceXAI, Elon Musk's company—is priced at $2 per million input tokens and $6 per million output tokens. Elon Musk responded to Zuckerberg's announcement on X with a single word: "Jinx"


On the other hand, Anthropic and OpenAI charge even more. The price for Anthropic's Claude Opus 4.

8 is $5 (approximately 480 rupees) per million input tokens and $25 (approximately 2,390 rupees) per million output tokens, respectively; meanwhile, OpenAI's GPT 5.6 Terra is priced at $2.5 (approximately 240 rupees) and $15 (approximately 1,430 rupees), respectively.

The launch of Muse Spark 1.1 comes during a busy week for Meta and the AI sector in general. On Tuesday, the company released Muse Image—its new image generation model—via the Meta AI app, WhatsApp, and Instagram.

Meanwhile, OpenAI has also made its GPT-5.6 model family available to the public and launched its AI agent for the workplace, ChatGPT Work.

This launch also marks a shift in Meta's AI strategy. Although the company had previously focused on open-source releases with its Llama family, it is now charging developers to use a proprietary model.

Alexandr Wang clarified that Meta remains "committed to open source" and that an open-source variant of Muse Spark is being developed, though he did not specify when it would be released. Wang added that Meta is training a more powerful model codenamed Watermelon.

Kahekashan is a passionate technophile with a keen eye for cutting-edge gadgets, emerging technologies, and everything in the digital realm. Raised in a Defence family with strong values and a background in literature, she has consistently pursued excellence in every endeavour. Her last full-time assignment involved content writing with the Indian School of Business.

Next Story
Share it