so there's no flash<br>of the wrong palette before React hydrates. Mirrors the useTheme hook<br>(storage key "rw-theme"; classes rw-theme-dark|light|ascii). -->

Futurism says Meta has “almost nothing” to show for AI. The numbers say otherwise - RuntimeWire

RuntimeWire

You're browsing RuntimeWire with JavaScript disabled. Articles and<br>navigation work fully. Interactive features — search, comments,<br>and newsletter signup — require JavaScript.

Why it matters

Meta's models and distribution make Zuckerberg's AI strategy credible. Its $130B-$145B capex plan still demands returns far beyond the early usage Meta has disclosed.

Futurism wanted a corpse. Meta gave it a messy earnings report, so the publication wrote an obituary.

“Despite having almost nothing to show from his enormous spending on AI,” begins the July 31 article, headlined “Mark Zuckerberg’s Pivot to AI Is Blowing Up in His Face Spectacularly.”

By the end, Meta is described as “practically absent from the frontier AI model race.” Muse Spark is supposedly outdone by competitors in “most tasks.” Muse Image is waved away as an “afterthought.” Meta’s users and engagement are said to “continue to slide.”

Almost every part of that picture collapses under inspection.

The article does technically mention Meta’s newest models. That is the trick. It gives each one enough space to dismiss it, while withholding the benchmarks, pricing, adoption and distribution that would make the story considerably harder to write.

Muse Spark is already a frontier model

Muse Spark 1.1 trails the leading models in some long-running agentic tasks. Meta has acknowledged those gaps, and Axios reported that agents from OpenAI, Anthropic and Google can currently operate for longer and handle a wider variety of jobs.

Across the whole model, the results look very different from Futurism’s description.

Independent evaluator Artificial Analysis scored Muse Spark 1.1 at 51 on its Intelligence Index, effectively tied with GPT-5.4 at xhigh reasoning, GPT-5.6 Luna and GLM-5.2. The model scored 45% on Humanity’s Last Exam, one point behind Claude Opus 4.8. It ranked third among all tested models on SciCode.

Muse Spark also carries a one-million-token context window and completed Artificial Analysis’s evaluation with fewer output tokens than the similarly scoring GPT-5.4, GPT-5.6 Luna and GLM-5.2. At Meta’s API prices, Artificial Analysis estimated a cost of roughly 26 cents per Intelligence Index task.

That is a credible, economically competitive frontier model with specific weaknesses. Artificial Analysis published the complete results on July 10.

Futurism published three weeks later and included none of them.

Calling Muse Image an “afterthought” is indefensible

Futurism devoted one sentence to Muse Image:

“At the same time, Meta released an image-generation model called Muse Image, which felt like an afterthought and a too-little-too-late attempt to catch up with its competitors.”

“Felt like” is doing tremendous work there.

At launch, Meta reported that Muse Image held the No. 2 position on Arena’s human-preference rankings in three separate categories: text-to-image generation, single-image editing and multi-image editing. Its forthcoming Muse Video model ranked third for text-to-video.

Muse Image can search for visual references, write and execute code when a request requires precision, revise its own output, preserve coherence across editing turns and combine people, objects, clothing and environments from multiple reference images. It works jointly with Muse Spark on tasks such as building websites and animated media.

Those rankings came through Meta’s announcement and deserved outside scrutiny. A skeptical reporter could have examined the evaluation methodology, tested the product or found evidence that Arena performance does not translate into consumer adoption.

Futurism supplied an adjective.

The underlying technical announcement contained enough material to challenge Meta seriously. The article gave readers none of it.

The Axios citation does not say what Futurism implies

Futurism writes that Muse Spark is “still easily outdone in most tasks by competing models from OpenAI, Anthropic, and Google,” attributing that judgment to Axios.

The linked Axios report says rival agents can handle a broader range of tasks and operate autonomously for longer. That is a narrower and more defensible observation about one area of model performance.

Another Axios report on the Muse Spark 1.1 launch noted its improved coding and long-context abilities, competitive API pricing and deployment across Facebook, Instagram, WhatsApp and Meta AI.

Axios also identified the strategic advantage missing from Futurism’s story. Meta AI chief Alexandr Wang described the billions of people using Meta’s products, along with the company’s knowledge about those users, as something competitors cannot replicate.

Turning a weakness in extended agent operation into “easily outdone in most tasks”...