{"slug": "meta-launches-low-cost-muse-spark-1-1-as-enterprise-ai-spending-comes-under", "title": "Meta launches low-cost Muse Spark 1.1 as enterprise AI spending comes under scrutiny", "summary": "Meta launched Muse Spark 1.1, a frontier AI model that rivals leading LLMs on coding and agentic benchmarks while undercutting OpenAI and Anthropic on API pricing, potentially reshaping enterprise AI procurement as inference costs come under scrutiny.", "body_md": "Meta has unveiled Muse Spark 1.1, saying the frontier AI model rivals leading LLMs on coding, computer use, and agentic AI benchmarks while undercutting OpenAI and Anthropic on API pricing, potentially lowering the cost of deploying AI agents in enterprises.\n\nMeta unveiled Muse Spark 1.1 on Thursday, pairing frontier-model performance with aggressive pricing in a move that analysts say could pressure rivals such as OpenAI and Anthropic and reshape enterprise AI procurement decisions.\n\nMeta is betting that lower inference costs can help it gain ground in the enterprise AI market with the launch of Muse Spark 1.1, a frontier model that rivals top competitors on key benchmarks while costing a fraction as much to deploy.\n\nThe latest model, which was [teased](https://www.infoworld.com/article/4192724/metas-ai-chief-says-new-muse-spark-update-will-sharpen-coding-agentic-ai.html) last week, matched or was competitive with leading models, such as Claude Opus 4.8, Gemini 3.1 Pro, and GPT 5.5, across several agentic AI, coding, and computer-use benchmarks, including SWE-bench Verified, Terminal-bench, BrowseComp, SpreadsheetBench, and OSWorld, Meta wrote in a blog [post](https://ai.meta.com/blog/introducing-muse-spark-meta-model-api/).\n\nMuse Spark 1.1, which is currently in public preview and available via the Meta Model API, will cost $1.25 per million input tokens and $4.25 per million output tokens, the company [noted](https://developer.meta.com/ai/products/meta-model-api/).\n\nBy comparison, OpenAI [charges](https://developers.openai.com/api/docs/pricing) $5 per million input tokens and $30 per million output tokens for GPT-5.5, while Anthropic [charges](https://platform.claude.com/docs/en/about-claude/pricing) $5 and $25, respectively, for Claude Opus 4.8. Google’s Gemini 3.1 Pro, on the other hand, is [priced](https://ai.google.dev/gemini-api/docs/pricing) at $2 per million input tokens and $12 per million output tokens.\n\nThat sheer difference in API pricing, according to [Pareekh Jain](https://pareekh.com/about/), principal analyst at Pareekh Consulting, is enough to attract CIOs’ attention, at least for pilots, at a time when enterprises are trying to scale agentic deployments: “Pricing matters because inference costs increase rapidly when thousands of agents are working continuously.”\n\n“Output tokens are often the largest model expense in coding, customer service, and process automation agents. Muse Spark’s output price is about 86% below GPT-5.5 and more than 90% below Claude Opus 4.8,” Jain said.\n\nHowever, [Muskan Bandta](https://www.linkedin.com/in/muskan-bandta2004), cloud associate at FinOps services providing firm ZopDev, pointed out that the price is not a guarantee of adoption, despite the fact that most enterprises are likely to deploy the Muse Spark 1.1 for new projects.\n\n“Cost becomes the primary differentiator only once the model is judged good enough. Developers don’t pick the cheapest model; they pick the cheapest model that clears their quality bar. So, price is the reason people show up, capability is the reason they stay,” Bandta said.\n\nSimilarly, CIOs are also likely to put more emphasis on the model’s security, data protection, uptime, audit trails, regional availability, support, and predictable behavior, rather than just the price, Jain said.\n\nThat distinction, according to Bandta, reflects a familiar pattern in enterprise technology buying: “This is the same lesson we saw in the cloud, where the cheapest provider on paper rarely won the biggest enterprise share. Price is one input in the total cost of ownership that includes risk, control, and switching cost, not the whole decision.”\n\nEven so, the lower pricing could still shift the balance of power in enterprise procurement, Jain said: “This could help CIOs negotiate larger volume discounts, committed-use agreements, and better pricing from OpenAI, Anthropic, and cloud providers. It also strengthens the case for multi-model procurement rather than depending on one vendor.”\n\n“Companies that do not even adopt Muse Spark can also use its pricing as evidence that frontier-level inference is becoming cheaper,” Jain added.\n\nAnalysts pointed out that Meta’s new model could intensify competition in the frontier model market by forcing rivals to compete on inference economics and model sizes.\n\n“It’s a real shot across the bow, and I’d expect OpenAI and Anthropic to respond on two fronts. Some of it will be price, cheaper tiers, and better cached and batch rates, because Meta has just reset what the market thinks a frontier token should cost,” Bandta said.\n\n“But the incumbents won’t win the race with lower-priced offerings and more flexible pricing models. I expect them to lean harder into the things price can’t buy, governance, security, reliability, and enterprise support, to justify premium pricing,” Bandta added, likening the shift to an “early innings” of a price war that the industry saw with the expansion of cloud.\n\n“The cloud infrastructure price war showed that while prices fell over time, vendors ultimately differentiated themselves through platform capabilities rather than cost alone,” Bandta further added.\n\nIn contrast, [Amit Jena](https://www.linkedin.com/in/znamit/), head of AI at IT consulting firm Kanerika, pointed out that a cloud-infrastructure-style pricing war was unlikely: “Frontier models are capital-intensive; margins are already thin. Vendors can’t sustain aggressive repricing without sacrificing quality.”\n\nRather, Jena sees Meta increasing prices soon after launch: “History suggests what happens next — aggressive entry pricing, then repricing once market share solidifies. See Meta’s advertising platform and cloud pricing evolution across the industry. If that pattern repeats, pricing could rise 30–50% in 18–24 months.”\n\nFor now, Meta is offering developers $20 in free API credits to experiment with Muse Spark 1.1.\n\n*The article originally appeared on InfoWorld.*", "url": "https://wpnews.pro/news/meta-launches-low-cost-muse-spark-1-1-as-enterprise-ai-spending-comes-under", "canonical_source": "https://www.computerworld.com/article/4195528/meta-launches-low-cost-muse-spark-1-1-as-enterprise-ai-spending-comes-under-scrutiny-2.html", "published_at": "2026-07-10 10:03:51+00:00", "updated_at": "2026-07-10 10:09:19.882161+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "ai-products", "ai-infrastructure", "ai-agents"], "entities": ["Meta", "Muse Spark 1.1", "OpenAI", "Anthropic", "Google", "Pareekh Jain", "Muskan Bandta", "ZopDev"], "alternates": {"html": "https://wpnews.pro/news/meta-launches-low-cost-muse-spark-1-1-as-enterprise-ai-spending-comes-under", "markdown": "https://wpnews.pro/news/meta-launches-low-cost-muse-spark-1-1-as-enterprise-ai-spending-comes-under.md", "text": "https://wpnews.pro/news/meta-launches-low-cost-muse-spark-1-1-as-enterprise-ai-spending-comes-under.txt", "jsonld": "https://wpnews.pro/news/meta-launches-low-cost-muse-spark-1-1-as-enterprise-ai-spending-comes-under.jsonld"}}