Skip to content
VECTOR WIREAI INTELLIGENCE
⌘K
UTC
Models739Deals30d257Policy30d64, 14 pendingSources1,010

Muse Spark

MetaReleased Apr 8, 2026Unpriced

Score basis

Muse Spark is capable in instruction following, factuality, long context, multimodal tasks, reasoning, and coding; and behind the leaders in agentic tasks. Too few results yet to rate safety, math, or multilingual tasks.

Capability profile

Bars show the model's median result as a share of the leading model's, per capability. Select a row to see its results.

Capable6 capabilities
  1. Instruction Following−14.0%2 of 3
  2. Factuality−14.7%3 of 4
  3. Long Context−14.9%1 of 3
  4. Multimodal−18.3%2 of 6
  5. Reasoning−19.3%4 of 6
  6. Coding−23.1%6 of 10
Limited1 capability
  1. Agentic−31.9%3 of 7
Not rated3 capabilities

Too few results yet to rate Safety, Math or Multilingual.

Price

No current price is tracked for this model. See the rate card

Evidence

46results on40benchmarks

  • 9 independently verified
  • 19 aggregator
  • 14 vendor-reported
  • 4 cross-referenced

From 12 sources · latest Oct 8, 2026 · How verification works

Research

3 papers reference Muse Spark

Muse Spark benchmark results

46 results on 40 benchmarks, grouped by capability. Each bar is the result as a share of the capability leader's; every result links to its source.

Instruction Following

CapableFull instruction following ranking

14.0% behind the leader2 of 3 ranked benchmarks measured

Factuality

CapableFull factuality ranking

14.7% behind the leader3 of 4 ranked benchmarks measured

Long Context

CapableFull long context ranking

14.9% behind the leader1 of 3 ranked benchmarks measured

Multimodal

CapableFull multimodal ranking

18.3% behind the leader2 of 6 ranked benchmarks measured

Show 3 more multimodal resultsHide 3 multimodal results

Reasoning

CapableFull reasoning ranking

19.3% behind the leader4 of 6 ranked benchmarks measured

Show 3 more reasoning resultsHide 3 reasoning results

Coding

CapableFull coding ranking

23.1% behind the leader6 of 10 ranked benchmarks measured

Show 1 more coding resultHide 1 coding result

Agentic

LimitedFull agentic ranking

31.9% behind the leader3 of 7 ranked benchmarks measured

Show 1 more agentic resultHide 1 agentic result

More results

Benchmarks outside the capability baskets. They are not ranked against a leader.

Show 11 more resultsHide 11 results

Muse Spark: common questions

Who makes Muse Spark?

Muse Spark is made by Meta.

When was Muse Spark released?

Muse Spark was released on Apr 8, 2026, according to Artificial Analysis.

What is Muse Spark good at?

Muse Spark is capable in instruction following, factuality, long context, multimodal tasks, reasoning, and coding; and behind the leaders in agentic tasks. Too few results yet to rate safety, math, or multilingual tasks.

How many benchmarks has Muse Spark been tested on?

We track 46 results for Muse Spark on 40 benchmarks from 12 sources, 9 of them independently verified. The latest was recorded on Oct 8, 2026.

About this record

Where Muse Spark's numbers come from, and every name it appears under.

Tracked since
Apr 25, 2026
Newest source mention
Jul 13, 2026

Where the results come from

Verification: 46 scores · 9 independently verified · 19 aggregator-attributed · 4 vendor cross-reference · 14 vendor-reported. How these tiers are assigned

From 12 sources on 7 sites. Artificial Analysis supplies 19 of them; the 9 independently verified results come from 4 sites. Bars are coloured by trust tier.

  • artificialanalysis.ai19
  • api.llm-stats.com13
  • ai.meta.com5
  • labs.scale.com3
  • datasets-server.huggingface.co2
  • epoch.ai2
  • lmarena.ai2

Also known as

muse spark*muse spark (contemplating mode)superintelligence labs muse sparkmuse spark 1.0