Meta’s Dirt-Cheap Muse Spark 1.1 Crashes the AI Coding Party and Threatens a Price War

• Reentering the AI race
• Public API and enterprise push
• The price undercut
• Competitive stakes and market reaction

Meta’s Dirt-Cheap Muse Spark 1.1 Crashes the AI Coding Party and Threatens a Price War
Meta’s latest AI coding model, Muse Spark 1.1, has entered an already crowded market with one clear weapon: price. The move raises a sharp question for rivals like OpenAI, Google, and Anthropic—compete on costs or risk losing enterprise developers.

§Reentering the AI race

Meta first reentered the coding-AI race in April with the initial Muse Spark model. On July 9, it unveiled Muse Spark 1.1, a multimodal, “agentic” system that can manage complex workflows, fix bugs, and handle large code migrations for enterprises. Meta describes the new release as a “step-change” over the first generation, citing better support for end-to-end agentic workflows, multi-agent systems, and native perception across images, video, and documents.

§Public API and enterprise push

Alongside the model, Meta launched the Meta Model API, giving US developers public-preview access to Muse Spark 1.1 for the first time. CEO Mark Zuckerberg framed the release as a push into long-running, tool-using software agents, calling Spark 1.1 “a strong agentic and coding model at a very low price.” He highlighted strengths in “agentic performance, tool use, and computer use” and promised “more to come soon.”

§The price undercut

Behind the marketing is an aggressive pricing play. Meta will charge $1.25 per million input tokens and $4.25 per million output tokens, significantly below Google and OpenAI’s flagship models. Internally, Meta pitches the model as delivering “frontier or very high-level intelligence at a much more affordable cost.”

External observers are already reading it as a margin-crushing move. One AI practitioner said the model is “so…cheap I almost don’t believe it,” claiming it runs at roughly one-tenth the cost of some leading rivals. Another benchmark-focused account touted Muse Spark 1.1 as new state of the art on several legal and tax-agent tests while being “10x cheaper and twice as fast” than a leading competitor.

§Competitive stakes and market reaction

Business analysts note that Muse Spark 1.1 is Meta’s first model it will directly charge for, potentially opening a major new revenue stream if it can displace popular coding tools from Anthropic, OpenAI, and others. Meta says enterprise customers like coding startup Cline see the pricing as enabling “heavy AI coding tasks at scale,” describing that combination of power and cost as “rare.”

Zuckerberg has publicly criticized competitors’ “very extreme” pricing and claims Spark 1.1 beats Google’s Gemini on several agent and coding benchmarks. With Meta also rolling out a companion image model, Muse Image, the company is signaling that it intends to battle on both capability and price—risking an industry-wide race to the bottom on margins even as demand for AI coding tools surges.

Continue reading https://foxvector.com/stories/019f604b-730b-2e23-7392-3729b42ff199