
On July 9 Meta released Muse Spark 1.1, a significant upgrade from the first Muse Spark — and, just as notably, launched the public preview of the new Meta Model API alongside it.
The pitch, from the top:
"Today we're releasing Muse Spark 1.1 — a strong agentic and coding model at a very low price. Available through our new Meta Model API and in Meta AI." — Mark Zuckerberg
Superintelligence Labs chief Alexandr Wang put a benchmark stake in the ground: Muse Spark 1.1 is "an industry-competitive agentic and coding model; across many agentic evals it rivals GPT-5.5 and Opus 4.8".
Why it matters
The model is only half the story. The Meta Model API is Meta's first real move to sell its models as a developer platform rather than just run them inside Meta AI — and it lands a week after Meta Compute, the company's entry into selling hosted models and raw GPU capacity. Together they sketch the same strategy from two sides: turn a $115–135B infrastructure bill into a revenue line, and compete for developers on price ("very low price" is doing deliberate work in that launch copy — this is the token-billing-shock month, and Meta knows it).
Skeptics will note the benchmarks are Meta's own and "many agentic evals" is carrying weight.
Sources: @AIatMeta announcement, Mark Zuckerberg, Alexandr Wang
About the Authors
Federico Ulfo
Founder, Engineer
New York City