# Meta to release open weights version of Muse Spark 'soon'

Zuckerberg announces smarter, more efficient version of the AI model.

- **Source:** [www.theregister.com - Articles](https://brifing.si/source/2c34ea50-d341-4a10-925d-12c012a9ad10)
- **Published:** 2026-09-02
- **Topic:** [Data & AI](https://brifing.si/c/data-ai)
- **Original article:** <https://www.theregister.com/ai-and-ml/2026/09/02/zucks-muse-to-spark-joy-with-open-weights-release-soon/5294093>
- **On Brifing:** https://brifing.si/s/1251a6a2-fcc2-406b-bf1a-23902ac921b9

## Open Source Announcement

Meta CEO Mark Zuckerberg announced on X that the company will soon release an open weights version of its flagship AI model Muse Spark. This move is seen as a significant step in the AI ecosystem. Since its initial release in April, Muse Spark has undergone several refinements, demonstrating Meta's commitment to continuous improvement in its AI offerings.

## New Version Features

Muse Spark 1.3 is introduced as smarter, more efficient, and less verbose compared to previous versions. The model shows significant improvements particularly when used with AI agents and coding assistants. According to Meta's blog post, the model now asks clarifying questions when prompts are ambiguous and seeks user confirmation before taking important actions.

## Performance Enhancements

The new version performs better in 'long-horizon' workloads and has more realistic capabilities assessment. Meta claims version 1.3 delivers solid gains across a wide range of benchmarks compared to 1.2. Independent benchmarking confirms a 4-point jump in overall intelligence score, putting it in close competition with leading models.

## Competitive Landscape

Independent tests had shown Muse Spark 1.2 performing on par with GPT 5.6 Terra and Z.AI's GLM 5.3 Flash. The new version now competes closely with GPT 5.6 Sol, Claude Opus 5, and Grok 4.6 High. This development highlights the intense competition in the AI space, with major players continuously pushing the boundaries of performance.

## Cost Advantage

Meta positions Muse Spark as a relatively affordable frontier model, with version 1.3 further improving this advantage. The contributor tier offers significant discounts, with prices as low as $0.002 (cached input), $0.10 (input), and $0.20 (output) per million tokens. This pricing strategy aims to increase accessibility to advanced AI capabilities.

## Strategic Moves

Meta's investments and frequent updates to Muse Spark are seen as efforts to maintain competitive pressure in the AI space. The company wants to demonstrate that its substantial capital expenditures are yielding results. The recent release of OpenClaw 2.0 adds to the industry's dynamic competitive environment.

## User Experience

Version 1.3 improves user experience by reducing the number of turns and tokens needed to complete tasks. The model's ability to seek human counsel in ambiguous situations and confirm before taking consequential actions aims to reduce error rates. These features make the model more practical for real-world applications.

## Industry Impact

The upcoming open weights release of Muse Spark could mark a significant milestone in the AI ecosystem. Meta's move may pressure other industry players to follow suit with similar initiatives. Such developments are expected to contribute to the democratization of AI technology in the coming years.

## Sources covering this story

- https://www.theregister.com/ai-and-ml/2026/09/02/zucks-muse-to-spark-joy-with-open-weights-release-soon/5294093 — 

---

_Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._
