Sourced model profile · Current through July 21, 2026

Muse Spark

A closed-weight Meta API model documented across multimodal, language, and vision domains for generation and question answering.

Developed byMeta

Released: April 8, 2026

What the evidence says

These facts come from the reviewed Epoch AI catalog artifact and its linked primary record. They describe the model. They do not rank it.

Developer [1]: Meta
Released [1]: April 8, 2026
Domains [1]: Multimodal · Language · Vision
Tasks [1]: Language modeling/generation · Question answering
Access [1]: API access
Weights [1]: Closed

Where it may fit

These are evidence-bounded screening suggestions, not performance claims or purchase recommendations.

Vision-and-language question answering

Consider it when one hosted Meta API workflow must span language and vision inputs for question answering.

Evidence fields: domains, tasks, access [1]

Meta-hosted language generation

Consider it for language generation delivered through Meta's managed API rather than downloadable Llama weights.

Evidence fields: tasks, access, openWeights [1]

Limitations

The record does not supply context window, price, latency, benchmark results, or safety evaluations. The separately pinned LMArena snapshot lists a Muse Spark 1.1 entry that does not exactly match this catalog record, so no rating is shown here.

Current standing

According to the LMArena snapshot published July 16, 2026, no exact match for this catalog record appears in the style-controlled text category, so this page reports no current rating for Muse Spark.

See the full leaderboard

Across the boards

Each card below is one named authority's own current view of Muse Spark. Model Gauntlet mirrors these sources with attribution and does not combine them into a single score or rank Muse Spark across them.

LMArena overall

Human preference, style-controlled text

LMArena rating: 1487.4
Board rank: #7
Reported interval: 1481.5 to 1493.2
Battles: 13,562