Meta's Muse Spark 1.1 launches at 86-90% below rival output prices as enterprises cut AI spend

Short summary

Meta released Muse Spark 1.1 at roughly 86-90% below GPT-5.5 and Claude Opus 4.8 output prices, just as enterprises are cracking down on runaway AI bills. Companies like Uber and Microsoft are already capping or switching tools after multi-million-dollar monthly costs. The era of ignoring inference cost is over—teams must route by difficulty, set budgets, and tie spend to measurable outcomes.

• •Meta's Muse Spark 1.1 undercuts GPT-5.5 and Claude Opus 4.8 by 86-90% on output token pricing
• •Enterprises are pulling back: Uber capped agentic coding at $1,500/person/month; Microsoft cut Claude Code licenses
• •Practical advice: route by task difficulty, set per-team budgets, and measure ROI not token volume

Generated with AI, which can make mistakes.