<div id="readability-page-1"><div id="ArticleBody-InlineImage-108317329"><div><div><p>Alexandr Wang, chief AI officer of Meta, during the Bloomberg Tech conference in San Francisco, California, US, on Thursday, June 4, 2026.</p><p>David Paul Morris | Bloomberg | Getty Images</p></div></div></div><div><p><br/>Three months after unveiling its first artificial intelligence model under the leadership of AI chief Alexandr Wang, <span id="SpecialReportArticle-QuoteInBody-2">Meta<span><span id="-WatchlistDropdown"></span></span></span> is rolling out a major update as it attempts to compete with OpenAI and Anthropic in critical areas of the market. </p><p>Muse Spark 1.1, which Meta introduced on Thursday, represents its &#34;strongest model for agentic and coding work yet,&#34; Wang said in an interview with CNBC. The initial Muse Spark model released in April was only available to &#34;select partners&#34; who could access the technology via a &#34;private API preview.&#34;</p><p>Meta is making the new model's API available through a developer portal as part of a public preview, where users will be able to sign up and see instructions for integration. A Meta spokesperson said some early partners can already access the API, and new users &#34;will be able to add themselves to a waitlist and be added from there over time.&#34; For now, Meta said it's limiting API access to its own properties rather than making it available on third-party platforms like the popular OpenRouter marketplace.</p><p>&#34;This is going to be served on top of the computer infrastructure that we've built,&#34; Wang said.</p><p>It's Meta's second notable rollout for the Muse family this week. On Tuesday, Meta released Muse Image, originally code-named Mango, a model for creating images, as the company seeks to attract creators and advertisers to its offerings.</p><p>Meta CEO Mark Zuckerberg is coming under pressure from Wall Street to show a return on the company's massive and growing investment in AI infrastructure and development. While it's spending at the rate of its hyperscaler peers, Meta doesn't have a cloud infrastructure business (though it plans to start one), and it's failed to keep up with OpenAI, Anthropic and <span id="SpecialReportArticle-QuoteInBody-8">Google<span><span id="-WatchlistDropdown"></span></span></span> in developing popular models and AI applications. </p></div><div><p>Wang characterized pricing of the Muse Spark update as &#34;very aggressive and attractive&#34; compared with similar offerings from labs like Anthropic and OpenAI. He said every new API account will start with $20 in free credits. From there, the company will charge $1.25 per million tokens in input, and $4.25 per million tokens of output, he said. </p><p>&#34;The goal is to really have attractive pricing that scales with immense consumption usage,&#34; Wang said.</p><p>He said Muse Spark 1.1 outperformed rival models in certain tasks involving the ability to interact with various third-party coding products and services. </p><p>Wang's Meta Superintelligence Labs, or MSL, trained Muse Spark 1.1 to excel in coding-related tasks because that ultimately improves the capabilities of AI agents that can autonomously perform multiple tasks like a fleet of human interns, he said. </p><p>&#34;You kind of have to build coding capabilities as part of that in service of overall agentic capabilities,&#34; Wang said.</p><p>The tech industry's excitement about AI agents took off in the first half of 2026, in part due to the sudden popularity of OpenClaw, which developers could use to manage AI models that power supercharged digital assistants. Wang said Meta trained Muse Spark 1.1 &#34;to be able to work well with all of the most popular harnesses that developers use today, and we felt that was the best approach for this model given our goal to maximize adoption.&#34;</p><p>Although Meta's previous AI strategy emphasized releasing its earlier Llama family of models to the open-source community, the company is now focusing on selling access to proprietary AI models.</p><p>Wang said that Meta is still &#34;committed to open source&#34; and that his MSL unit has a &#34;variant of Muse Spark that is in development that we do intend to open source.&#34; He declined to say when the company would release it. </p><div><p>Wang added that he's been &#34;dog-fooding&#34; the latest Muse Spark model, and is excited about the technology's ability to be used as tool for improving personal health via tasks like searching the web, reading academic papers and accessing personal health-related data.</p><p>&#34;It's one of these use cases that I think really encapsulates the needs of these agentic systems,&#34; Wang said of his AI and health experiments.</p></div><p>Wang said Meta is currently training a more powerful AI model, code-named Watermelon, but didn't say when it would be released. Muse Spark's code name was Avocado. </p><p><strong>WATCH</strong>: One year into Meta Superintelligence Labs.</p></div><div><p>Choose CNBC as your preferred source on Google and never miss a moment from the most trusted name in business news.</p></div></div>

METU

METW

METD

ANTH

FBYY

MAGX

<p>Meta launched Muse Spark 1.1, an upgraded AI coding model led by Alexandr Wang, to compete with OpenAI and Anthropic. The update offers aggressive pricing and public API access via a developer portal, marking a shift towards proprietary models while maintaining open-source commitments. This rollout aims to enhance agentic capabilities and address Wall Street pressure for AI returns.</p>

- Meta has introduced a major update to its AI model, Muse Spark 1.1, to enhance competitiveness against OpenAI and Anthropic.  
- The new model's API will be publicly available through a developer portal, with early partners having access and new users able to join a waitlist.  
- CEO Mark Zuckerberg faces pressure to demonstrate returns on AI investments, while pricing for Muse Spark is positioned as competitive in the market.

Meta jumps into AI coding market in effort to chase Anthropic and OpenAI