Meta CEO Mark Zuckerberg announced on November 9 (US time) that Muse Ember, Muse Spark 1.3, and Muse Wonder are now available on Google Cloud, and Sundar Pichai replied "welcome to the model garden." Google's official docs list all three as Preview partner models, served only on the global endpoint through a stateless OpenAI-compatible Responses API. Google's page does not list prices, so Meta's own rate card remains the reference.
Key Takeaways
- ✓Scope: 3 models, Muse Wonder 1.0, Muse Spark 1.3, and Muse Ember 1.0, all marked Preview in Google's partner model list (Google Cloud docs)
- ✓Access: a single global endpoint with the stateless OpenAI-compatible Responses API;
store: trueis not supported; pass the bare model IDmuse-wonder-1.0,muse-ember-1.0, ormuse-spark-1.3 - ✓Pricing reference (Meta's official rate card; Google's page lists no prices): Wonder $4 / $20, Spark 1.3 $1.25 / $4.25, Ember $0.10 / $0.50 per million input/output tokens
- ✓Official positioning: Wonder handles long-horizon coding and agent planning, Ember handles high-volume tool calls, routing, and extraction, and Google's docs suggest pairing Ember with Wonder or Spark 1.3; Spark 1.3 supports MCP and a 1M context window
- ✓Zuckerberg's post drew 5,300+ likes and about 250K views; Sundar Pichai's "welcome to the model garden" reply drew 2,100+ likes; LiteLLM already has PR #50768 adding support
Key Decision Metrics at a Glance
Turn your technical choice into a development budget
Compare 40 dev plans & simulate token costs vs $20/mo subscriptions
Project Links & Resources
Direct AccessIn-Depth Technical Analysis
Core Background and Industry Pain Points
Meta launched Muse Ember, its smallest model, on November 9, and the same day announced that Ember, Muse Spark 1.3, and its October flagship Muse Wonder are available on Google Cloud. Until now these models were only reachable through Meta's own Model API and Muse Code. For enterprises already on Google Cloud, the main barrier is often not model quality but signing another vendor and passing another security review.
Architecture Highlights and Underlying Mechanisms
According to Google's official docs, the three models are offered as managed APIs on the Gemini Enterprise Agent Platform and are enabled from their Model Garden model cards. Requests go to publishers/meta/v1/responses on the global endpoint, using a stateless OpenAI-compatible Responses API with SSE streaming; store: true is not supported. The docs note that Meta models are not a Google product and fall under the "Separate Offerings" terms.
Authoritative Benchmarks and Measured Comparisons
This is a distribution change; the models themselves are unchanged, and Google has published no new benchmarks. Official or vendor-reported references: Ember scores 36 on the Artificial Analysis Intelligence Index, tied with Claude Haiku 5.5, and Wonder previously scored 46.1% on CursorBench. Google's page lists no prices, so costs can only be estimated from Meta's rate card for now.
Developer Adoption and Getting Started
Find and enable the model card in Model Garden, get an access token with gcloud, and POST to the global endpoint with the bare model ID in the model field. Because only a global endpoint exists, teams with data residency requirements should check compliance first. The API is stateless, so multi-turn apps must keep their own conversation state. The official guidance is to plan with Wonder or Spark 1.3 and hand high-volume tool calls to Ember.
Benchmark side-by-side against alternatives, or calculate monthly token cost vs subscription break-even.
Discussion & Comments
0Sign in to join the discussion
Connect with AI developers to exchange benchmark insights.