/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

days · browse · Enter similar · o open

Meta's Muse Spark 1.2 scores 54 on the Artificial Analysis Intelligence Index, putting Meta next to SpaceXAI in a tie for third place amongst US labs

Artificial Analysis

Context & Ripple Effects

Meta introduced Muse Spark through Meta Superintelligence Labs as a model for Meta AI, including shopping-related queries, and said it intended to release a version under an open-source license. Its later agentic Meta AI update extended the model family into Gmail- and Google Calendar-connected tasks.

The 1.2 score gives that product rollout an independent ranking marker: Meta now shares third place among US labs with SpaceXAI, rather than asking users and developers to assess its model solely through product claims.

First-order effects

  • Meta gains a 54-point Artificial Analysis Intelligence Index result and a tie with SpaceXAI for third among US labs, strengthening the competitive positioning of Muse Spark 1.2.
  • Meta AI teams can point to a third-party benchmark alongside the model's existing deployment for Meta AI queries and shopping mode.

Second-order effects

  • The planned open-source Muse Spark release gains a more concrete performance reference for developers evaluating whether to adopt Meta's model family.
  • SpaceXAI now shares its third-place US-lab position with Meta, making subsequent benchmark updates a visible measure of separation between the two labs.

Third-order effects

  • If Meta continues pairing competitive benchmark results with deployment across Meta AI and connected-agent features, model competition will increasingly turn on both measured capability and distribution into existing consumer products.

The trend: Frontier AI labs are using independent benchmark standing to reinforce model launches that are already embedded in large product surfaces and developer ecosystems.

Discussion

  • @andrew_n_carr Andrew Carr on x
    I've seen a lot of people grumbling about muse spark 1.2 contributor slurping up your coding data, but didn't we all just go through that happily with Cursor -> grok 4.5 Seems odd to complain now (not that they're necessarily wrong, I guess).
  • @jsrailton John Scott-Railton on x
    Core asymmetry in big AI: every proprietary bit of knowledge you share is a training signal accruing to the API provider. You get none of it back. Thing is, @Meta built a fix. Private Processing for @WhatsApp serves inference in a Trusted Execution Environment (TEE) where even
  • @rihardjarc Rihard Jarc on x
    A few thoughts on the $META AI model's progress, because I think it is significant.  1. It does seem that $META has now leapfrogged $GOOGL in model quality when it comes to Muse 1.2 for many use cases, which is very surprising given the timeframe.  2. This is still the “Muse Spar…
  • @millionint Jerry Tworek on x
    When I heard some time ago that Meta is the third best AI lab right now, I was skeptical. But every day passing has been reinforcing that it's actually true. TBD strategy has succeeded and congratulations to the team that made it! The world needs more successful AI labs
  • @valsai @valsai on x
    Muse Spark 1.2 just cracked the top 5 on the Vals Index, at just $0.69 per test. This is 3x cheaper than Kimi and 10x or more cheaper than Fable, Opus, and 5.6 Sol. [image]
  • @chamath Chamath Palihapitiya on x
    Tactical Game Theory: Meta Scorched Earth This should have been Meta's play two years ago. That said, they are in an even better position to do it now considering the power and compute constraints that are emerging.
  • @artificialanlys @artificialanlys on x
    Meta has released Muse Spark 1.2.  It's their third release in four months and scores 54 on the Artificial Analysis Intelligence Index, significantly improving agentic knowledge work capabilities over prior releases and putting Meta next to SpaceXAI in a tie for third place among…
  • @louszbd Lou on x
    Looks like Muse Spark 1.2 made strong gains on long-horizon knowledge work. Impressive progress from the Meta team! [image]
  • @kunchenguid Kun Chen on x
    muse spark 1.2 look incredible - more to share soon once i played with it more but i want to highlight one thing first - this table basically tells you how valuable our usage data is.  same reason why anthropic and openai have been subsidizing their customer subscriptions it's ve…
  • @valsai @valsai on x
    Compared to Muse Spark 1.1, v1.2 improved 3.5 points overall (71.9 vs 68.4), led by a 10 pp gain on VibeCodeBench. That's where the price gap is starkest: $1.49 per test vs $41.74 for Claude Fable 5, roughly 28x cheaper and 3x faster for 10.6 pp lower accuracy.
  • @cline @cline on x
    We tried using Meta's new Muse Code agent, but it has a bug that doesn't let it sign in from a docker container.  So we did a fun experiment: Meta claims Muse Spark 1.2 was co-trained with their Muse agent harness.  So we extracted instructions from their system prompt and added …
  • @zephyr_z9 @zephyr_z9 on x
    Meta started a huge LLM price war