AIToday
Large Language ModelsAI Business & IndustryAI Stocks & MarketsSiliconANGLE AIPublished: Sep 3, 2026, 13:00 JST2 min read

Meta says Muse Spark 1.3 catches up with OpenAI, Anthropic

Meta says Muse Spark 1.3 catches up with OpenAI, Anthropic

Key takeaway

  • Meta claims its new Muse Spark 1.3 model is competitive with Anthropic and OpenAI.

  • An independent index ranks it third behind two Anthropic models.

  • Meta now charges for access to this model.

3 Key Points

  1. What happened

    Meta Platforms Inc. released Muse Spark 1.3, which it says puts it at the same level as the most recent models from OpenAI Group PBC and Anthropic PBC. The model scored 62 on an independent Artificial Analysis Intelligence Index, behind only Claude Fable 5.1 and Claude Opus 5 and ahead of OpenAI's models.

  2. Why it matters

    Meta has spent hundreds of billions of dollars on AI infrastructure, and investors are nervous about returns. The company has shifted from open-source Llama models to a proprietary strategy similar to Anthropic and OpenAI, and this release suggests Meta could finally start seeing payback from its AI investments.

  3. What to watch

    Developers will pay the same price for Muse Spark 1.3 as for version 1.2, launched in August. Meta has not decided whether to release Muse Spark 1.3's weights, and its next model, Watermelon, remains a work in progress.

Ask the AI about this article →

Context & Analysis

Meta's aggressive release cadence — four Muse Spark models in five months — shows how quickly the company is trying to close the gap with its rivals. The shift away from open-source weight releases marks a strategic pivot, as Meta now follows the proprietary path of Anthropic and OpenAI, partly to address investor concerns over massive spending. Wang's comments highlight efficiency gains, such as using roughly 25% fewer tokens for the same tasks, which could appeal to developers watching costs. The pending decision on weights and the future Watermelon model remain open questions that could shape Meta's competitive position.

FAQ

How much does Muse Spark 1.3 cost for developers?
According to Meta Chief AI Officer Alexandr Wang, developers will not pay more for Muse Spark 1.3 than they did for version 1.2, which launched in August.
Will Meta release the weights for Muse Spark 1.3?
Meta has not yet decided whether to release Muse Spark 1.3's weights. The company previously promised to release weights for 1.2 but has not done so yet.
How does Muse Spark 1.3 compare to Anthropic and OpenAI models?
Wang said it is "competitive" with Claude Fable 5.1 and "better than" OpenAI's GPT-5.6 Sol model. An independent analysis put its score at 62, behind only Fable 5.1 and Claude Opus 5.
SiliconANGLE AIRead Original Article

Get the latest Large Language Models news every morning

For example, today's edition would include:

  • Meta's Muse Spark 1.3 matches GPT-5.6-Sol, claims #3 modelLatent Space · 3h ago
  • CABiNet vs YOLO26-sem on UAVid: Accuracy, Compute, and GPU Latencyr/MachineLearning · 3h ago
  • C++ PCN library nears backprop accuracyr/MachineLearning · 3h ago

AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.

Free · takes 30 seconds · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. Q&As are published on this page for other readers too.

Related Articles

Next articleTSMC doubles down on equipment purchase as AI demand outstrips planning