Skip to content
VECTOR WIREAI INTELLIGENCE
⌘K
UTC
Models741Deals30d270Policy30d65, 13 pendingSources1,015

Muse Spark 1.2

MetaReleased Aug 5, 2026Current

Score basis

Muse Spark 1.2 is strong in factuality; capable in reasoning, long context, and agentic tasks; and behind the leaders in coding, instruction following, and math. Too few results yet to rate safety, multimodal tasks, 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.

Strong1 capability
  1. Factuality−5.9%3 of 4
Capable3 capabilities
  1. Reasoning−12.3%5 of 6
  2. Long Context−13.2%1 of 3
  3. Agentic−23.3%3 of 7
Limited3 capabilities
  1. Coding−26.1%5 of 10
  2. Instruction Following−31.8%1 of 3
  3. Math−37.1%1 of 5
Not rated3 capabilities

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

Price

$1.25input$4.25outputper million tokens

From Artificial Analysis · 3 providers tracked · All prices

Costs more than 71% of 330 priced models · 3:1 input-to-output blend, log scale

Evidence

34results on31benchmarks

  • 12 independently verified
  • 15 aggregator
  • 2 vendor-reported
  • 5 cross-referenced

From 9 sources · latest Oct 8, 2026 · How verification works

API features

Tool callingStructured outputsReasoning

As listed by OpenRouter

Muse Spark 1.2 benchmark results

34 results on 31 benchmarks, grouped by capability. Each bar is the result as a share of the capability leader's; every result links to its source.

Factuality

StrongFull factuality ranking

5.9% behind the leader3 of 4 ranked benchmarks measured

Reasoning

CapableFull reasoning ranking

12.3% behind the leader5 of 6 ranked benchmarks measured

Long Context

CapableFull long context ranking

13.2% behind the leader1 of 3 ranked benchmarks measured

Agentic

CapableFull agentic ranking

23.3% behind the leader3 of 7 ranked benchmarks measured

Coding

LimitedFull coding ranking

26.1% behind the leader5 of 10 ranked benchmarks measured

Show 1 more coding resultHide 1 coding result

Instruction Following

LimitedFull instruction following ranking

31.8% behind the leader1 of 3 ranked benchmarks measured

Math

LimitedFull math ranking

37.1% behind the leader1 of 5 ranked benchmarks measured

Multimodal

Not enough dataFull multimodal ranking

0 of 6 ranked benchmarks measured

More results

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

Show 7 more resultsHide 7 results

Muse Spark 1.2: common questions

Who makes Muse Spark 1.2?

Muse Spark 1.2 is made by Meta.

When was Muse Spark 1.2 released?

Muse Spark 1.2 was released on Aug 5, 2026, according to Artificial Analysis.

What is Muse Spark 1.2 good at?

Muse Spark 1.2 is strong in factuality; capable in reasoning, long context, and agentic tasks; and behind the leaders in coding, instruction following, and math. Too few results yet to rate safety, multimodal tasks, or multilingual tasks.

How much does Muse Spark 1.2 cost?

Muse Spark 1.2 costs $1.25 per million input tokens and $4.25 per million output tokens, according to Artificial Analysis. We track its price at 3 providers. At a mix of three input tokens to one output token, it costs more than 71% of the 330 priced models we track.

How many benchmarks has Muse Spark 1.2 been tested on?

We track 34 results for Muse Spark 1.2 on 31 benchmarks from 9 sources, 12 of them independently verified. The latest was recorded on Oct 8, 2026.

Which API features does Muse Spark 1.2 support?

OpenRouter lists tool calling, structured outputs, and reasoning for Muse Spark 1.2.

About this record

Where Muse Spark 1.2's numbers come from, and every name it appears under.

Tracked since
Aug 6, 2026
Newest source mention
Sep 26, 2026

Where the results come from

Verification: 34 scores · 12 independently verified · 15 aggregator-attributed · 5 vendor cross-reference · 2 vendor-reported. How these tiers are assigned

From 9 sources on 8 sites. Artificial Analysis supplies 15 of them; the 12 independently verified results come from 5 sites. Bars are coloured by trust tier.

  • artificialanalysis.ai15
  • livebench.ai7
  • deepmind.google5
  • api.llm-stats.com2
  • datasets-server.huggingface.co2
  • epoch.ai1
  • lmarena.ai1
  • simple-bench.com1

Also known as

muse-spark-1-2muse spark 1.2 (xhigh)muse-spark-1.2-contributormuse spark contributormusemeta muse spark 1.2