Meta Connect 2026: Muse Watermelon, Open Weights - Preview

Published September 23, 2026 - Menlo Park, California. Meta Connect 2026 starts Wednesday, September 23 with Mark Zuckerberg's keynote.

Meta's developer platform already has placeholder sections for Video Playground, Voice Playground, and Connectors (Meta developer portal, September 2026).

At a glance

• Meta Connect 2026 keynote: Wednesday, September 23, 2026.
• Expected reveals: Watermelon model family, Muse Spark 1.2 / 1.3 open weights, Muse Video Playground, Muse Voice Playground, and Stripe-payments-enabled Connectors.
• Muse Spark 1.3 (shipped September 2, 2026) matches Claude Fable 5.1 and beats GPT-5.6 Sol on coding benchmarks.
• Watermelon is Meta Superintelligence Labs' larger model in development - "extremely competitive" with frontier per Alexandr Wang, no public release date.
• Zuckerberg publicly promised Muse Spark open weights; Meta still intends to release the 1.2 weights; 1.3 open weights unconfirmed.

Data last verified September 23, 2026 from Meta AI Research, Meta developer portal, NYT reporting on the August 10 Watermelon framing, and TestingCatalog developer traces.

Quick Answer

Meta Connect 2026 starts Wednesday, September 23, 2026, and the keynote is expected to unveil the Watermelon model family, Muse Spark open weights, and a broader Muse developer platform.

Watermelon is Meta Superintelligence Labs' larger frontier model. Chief AI Officer Alexandr Wang has called it "extremely competitive" but has not yet dated the release (Meta, September 2026).

What is the Muse family

Source: Meta AI Research blog and developer portal (September 2026).

What we expect from the keynote

Three concrete items are visible on Meta's developer platform ahead of the keynote: Video Playground, Voice Playground, and Connectors (Notion, Granola, Stripe payments).

Meta recently expanded Model API with Muse Spark 1.3, Muse Image, and Muse Voice Transcribe. So the developer platform is already shipping pieces (TestingCatalog, September 22, 2026).

The rumored headline is the Watermelon model family. NYT reported on August 10 that Meta is building a "next-generation A.I. model, internally called Watermelon, that will be more powerful than Muse Spark."

Wang confirmed on the same day that "we believe that Watermelon will be extremely competitive" (NYT, August 10, 2026).

Zuckerberg echoed the framing in his September 2 Spark 1.3 tweet: "Next up 🍉 and Muse Spark open weights releases coming soon" (X, September 2, 2026).

How Muse Spark 1.3 benchmarks

Source: Meta AI Research benchmarks and Artificial Analysis (September 2026).

Why open weights matter here

Open weights let developers run the model on their own hardware, fine-tune it, and avoid per-token API spend.

Meta's Muse Spark 1.2 open-weights release is still pending final safety fine-tuning per Wang's September 3 statement. Muse Glimmer 30B (a 30B dense distillation) already shipped as open weights on August 9, 2026 (Meta AI Research, September 2026).

Meta's Contributor tier ($0.10 input / $0.20 output per million tokens, with training-data opt-in) is the cheapest frontier-class API in production. Muse Spark open weights extend the same playbook to the self-hosted market.

Anthropic, OpenAI, and Google do not currently offer a comparable tier or open-weights release (Meta, September 2026).

What the Connectors rollout means for developers

The Connectors launch lets third parties integrate with Muse through Stripe-payments-enabled API endpoints. Notion, Granola, and others are listed as launch partners.

The strategic angle is a developer platform play. Meta has historically lacked the third-party developer ecosystem that OpenAI has built through GPTs and the Assistants API. Connectors fills that gap with a revenue-sharing model.

For enterprise buyers, Connectors means Muse agents can pull live data from CRM, knowledge bases, and project tools without custom integration code. This closes the productivity-app gap that Microsoft Copilot and Google Workspace agents have already filled (Meta developer portal, September 2026).

For indie developers, the Stripe payment rail means a path to monetising Muse-built agents without standing up their own billing stack. The economics are still being published, but Meta's August 2026 developer blog described "favorable revenue share for Connectors partners."

Watch the keynote for the Connectors pricing and rate-card details. The platform launch is positioned as a complement to the Watermelon model reveal.

How this fits the frontier race

OpenAI prepared GPT-6 Sol and Luna models for launch in the same September 22-23 window. Meta's Watermelon timing lines up (TestingCatalog, September 22, 2026).

Anthropic shipped Opus 5.5 on September 22, 2026 and is testing Fable 5.2 ahead of release. Meta Muse Spark 1.3 already trades blows with Opus 5 on coding (Meta; Anthropic, September 2026).

xAI shipped Grok 4.7 on September 21, 2026. Google's Gemini 3.8 Flash shipped August 26 and targets consumer speed, not coding depth (xAI; Google, September 2026).

How to follow the keynote

The keynote is livestreamed at meta.com/connect and via Meta's developer YouTube channel. Meta will publish full session recordings and the developer API updates within 24 hours of the keynote.

The Muse Spark 1.2 open-weights release and the Watermelon model family announcement are the two highest-volume keywords to watch on the day (Meta, September 2026).

What the Watermelon family announcement could include

If Meta follows the pattern set by Muse Spark 1.3 in September 2026, expect three things from a Watermelon reveal.

First, the headline model announcement. A single frontier-class API model - probably a 70B+ dense or sparse mixture-of-experts architecture, tuned for coding and knowledge work. Pricing will likely mirror or undercut the Muse Spark 1.3 standard tier ($1.25/$4.25 per million).

Second, a multimodal expansion. Watermelon is positioned as a step beyond Muse Spark's text-and-image coverage. Expect native audio understanding (rivaling Google's Gemini audio), video input, and possibly native video generation via Muse Video.

Third, a developer-platform reveal. The Video Playground and Voice Playground sections on the developer portal are placeholders for Watermelon-era capabilities. Pricing for the Playground endpoints and the Connectors revenue share will likely be announced alongside the model.

If the keynote is a soft-launch, expect Watermelon weights behind a gated program rather than open release. Meta has been more cautious about open weights since the Muse 1.2 safety fine-tuning push slipped from August into September.

Why this matters for frontier competition

Meta's September 2026 push signals the company's confidence in the frontier race. Wang's August framing of Watermelon as extremely competitive, paired with the open-weights roadmap, is a direct response to OpenAI's closed-API strategy and Anthropic's pacing-the-frontier posture.

For enterprise buyers, the next four weeks resolve most of the open questions. Pricing, open weights availability, multimodal coverage, and tool-call benchmarks will all be updated across the major labs.

For self-hosted teams, Muse Spark 1.2 open weights are the single highest-impact event. Watch Meta's GitHub and Hugging Face accounts for the announcement.

Muse Glimmer 30B is already available as open weights - it is a 30B dense distillation that runs on consumer hardware. Teams evaluating the family should benchmark Muse Glimmer 30B first as a free reference, then test Muse Spark 1.3 API access.

The Contributor tier ($0.10 input / $0.20 output per million tokens, with training-data opt-in) remains the cheapest frontier-class API in production - and is a unique moat against OpenAI and Anthropic for teams comfortable sharing training data.

Internal links

For the Muse Spark 1.3 release details, see our Muse Spark 1.3 launch post.

For the IPO and capex backdrop driving the frontier race, see our OpenAI Anthropic IPO preparations post.

For a three-way comparison with Claude and Gemini, see our ChatGPT vs Claude vs Gemini guide.