Meta's Muse Spark 1.1 undercuts OpenAI, Anthropic by 86%

Key Takeaways

• Muse Spark 1.1 costs $1.25 per million input tokens versus $5 for GPT-5.5 and Claude Opus 4.8
• Output token pricing is 86% below OpenAI and 90% below Anthropic, critical for agent-heavy workloads
• Analysts expect CIOs to use Meta's pricing as leverage in negotiations, even without switching vendors

Meta has released Muse Spark 1.1, a frontier AI model priced at $1.25 per million input tokens and $4.25 per million output tokens. That is 75% cheaper than OpenAI's GPT-5.5 on input and 86% cheaper on output. The model, now in public preview via Meta's Model API, matched or came close to leading competitors on coding, computer use, and agentic AI benchmarks including SWE-bench Verified, Terminal-bench, and OSWorld.

The timing is deliberate. Enterprise CFOs have been pushing back on AI spending after months of experimentation without clear returns. Meta is betting that a model with competitive performance at a fraction of the cost will shift procurement conversations in its favor.

How does Muse Spark 1.1 pricing compare to competitors?

The numbers tell the story. OpenAI charges $5 per million input tokens and $30 per million output tokens for GPT-5.5. Anthropic charges $5 and $25 respectively for Claude Opus 4.8. Google's Gemini 3.1 Pro sits at $2 and $12. Meta's Muse Spark 1.1 undercuts all three on both metrics.

Output tokens matter most for enterprise deployments. Coding assistants, customer service agents, and process automation tools generate far more output than input. Pareekh Jain, principal analyst at Pareekh Consulting, put it plainly: "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."

Will CIOs actually switch to Muse Spark?

Price alone does not close deals. Muskan Bandta, cloud associate at FinOps firm ZopDev, made the distinction clear: "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."

CIOs evaluating Muse Spark will weigh security posture, data protection guarantees, uptime SLAs, audit trails, and regional availability. These concerns typically favor established vendors with enterprise sales teams and compliance certifications. Meta's enterprise AI infrastructure is younger than Microsoft's or Google's.

“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.”

— Muskan Bandta, ZopDev

Still, Jain sees indirect benefits even for companies that never touch Muse Spark: "Companies that do not even adopt Muse Spark can use its pricing as evidence that frontier-level inference is becoming cheaper." In procurement negotiations, that reference price is ammunition.

How will OpenAI and Anthropic respond?

Bandta expects a two-pronged response. "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," she said. "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."

This mirrors the early cloud wars. AWS, Azure, and Google Cloud competed on price for years, but enterprise adoption ultimately hinged on platform depth, compliance frameworks, and integration with existing toolchains. The same dynamic is emerging in LLM procurement.

Amit Jena, head of AI at consulting firm Kanerika, pushed back on the price war framing: "Frontier models are capital-intensive; margins are already thin. Vendors can't sustain aggressive pricing indefinitely." The implication: Meta may be buying market share now, but the industry won't race to zero.

What does this mean for multi-model strategies?

The real shift may be structural. Jain believes Muse Spark 1.1 strengthens the case for multi-model procurement rather than single-vendor lock-in. "This could help CIOs negotiate larger volume discounts, committed-use agreements, and better pricing from OpenAI, Anthropic, and cloud providers," he said.

Organizations running agentic workloads at scale, where thousands of agents run continuously, face inference costs that compound quickly. Having a credible low-cost option changes the math. Even if Muse Spark handles only a subset of tasks, it pressures the entire stack.

Frequently Asked Questions

Is Muse Spark 1.1 available now?

Yes, the model is in public preview and accessible via Meta's Model API.

How does Muse Spark 1.1 perform on coding benchmarks?

Meta claims it matched or was competitive with Claude Opus 4.8, Gemini 3.1 Pro, and GPT-5.5 on SWE-bench Verified and Terminal-bench.

Should enterprises switch from OpenAI or Anthropic to Meta?

Not automatically. Analysts recommend evaluating security, compliance, support, and regional availability alongside price and benchmark performance.

Will OpenAI and Anthropic lower their prices in response?

Analysts expect some price adjustments but believe incumbents will emphasize governance and enterprise support rather than match Meta's rates directly.

Source: Meta launches low-cost Muse Spark 1.1 as enterprise AI spending comes under scrutiny – Computerworld