Muse Wonder vs Muse Spark 1.1

  • Muse Wonder rates higher on LMArena’s text leaderboard: 1501 against 1478.
  • Muse Spark 1.1 is cheaper: $1.25 / $4.25 per million input / output tokens, against $4 / $20 for Muse Wonder.
  • On the 2 Vals AI benchmarks both have taken, Muse Wonder wins 2 and Muse Spark 1.1 wins 0.
LMArena text rating over time
14601470148014901500Jul 10Aug 5Aug 31Sep 26Oct 22
LatestMuse Spark 1.1 1478

Side by side

Models (2/6)
  • Muse Wonder
  • Muse Spark 1.1
BenchmarkMuse WonderMuse Spark 1.1
The model
LabMetaMeta
Released2026-10-122026-07-16
WeightsProprietaryProprietary
Context window1M1M
Input, per 1M tokens$4$1.25
Output, per 1M tokens$20$4.25
LMArena, people’s votes
LMArena TextMillions of people compare two anonymous models side by side and vote for the better answer1501#7 of 3511478#22 of 351
LMArena WebDevPeople ask two anonymous models to build the same web app, try both and vote for the better one1651#18 of 1241543#47 of 124
LMArena VisionSide-by-side votes on prompts that include an image—1293#26 of 146
LMArena CodingLMArena’s text votes on prompts that ask for code1518#11 of 3461498#28 of 346
LMArena MathLMArena’s text votes on math prompts1484#30 of 3391481#33 of 339
LMArena Creative writingLMArena’s text votes on stories, poems and other creative writing1493#7 of 3491435#57 of 349
LMArena Hard promptsLMArena’s text votes on the hardest prompts1505#12 of 3511485#27 of 351
LMArena Instruction followingLMArena’s text votes on prompts with specific instructions to follow1493#10 of 3511456#45 of 351
Vals AI, overall
Vals IndexVals AI’s headline score64.88%#6 of 27—
Vals AI, coding
Terminal-Bench 4.0Frontier-difficulty tasks an agent completes in a real terminal49.49%#10 of 55—
SWE-benchReal GitHub issues from open-source projects—82.00%#20 of 85
Vibe Code BenchBuilding complete web apps from a plain-language brief, graded on whether the app works83.09%#20 of 11572.16%#40 of 115
Code MigrationRewriting working programs in a new language, graded on whether the new program still passes its tests72.36%#5 of 28—
LiveCodeBenchFresh competitive programming problems, collected after models’ training cut-offs to avoid contamination—85.91%#29 of 143
Vals AI, finance
Finance AgentThe core tasks of a financial analyst, done as an agent with research tools65.55%#6 of 31—
Finance IndexVals AI’s composite of finance tasks61.23%#6 of 31—
Tax Agent BenchAgentic research into US corporate tax questions36.27%#12 of 76—
Vals AI, law
Legal IndexVals AI’s composite of legal tasks75.26%#3 of 28—
Legal Research BenchAgentic US legal research50.48%#6 of 8537.98%#39 of 85
Vals AI, reasoning
GPQA DiamondGraduate-level questions in biology, physics and chemistry that experts answer well and search does not—91.16%#23 of 138
MMLU-ProHarder, reasoning-focused multiple choice across 14 subjects, from law to engineering—88.73%#17 of 138
MMMUCollege-level questions that need both an image and text—86.59%#18 of 93
Terminal-Bench ScienceExpert-written scientific research workflows, run by an agent in a terminal22.86%#10 of 42—

Add models in the compare tool.

Questions

Which is better, Muse Wonder or Muse Spark 1.1?

Muse Wonder rates higher on LMArena’s text leaderboard: 1501 against 1478. Muse Spark 1.1 is cheaper: $1.25 / $4.25 per million input / output tokens, against $4 / $20 for Muse Wonder. On the 2 Vals AI benchmarks both have taken, Muse Wonder wins 2 and Muse Spark 1.1 wins 0.

Which is cheaper, Muse Wonder or Muse Spark 1.1?

Muse Spark 1.1 is cheaper: $1.25 / $4.25 per million input / output tokens, against $4 / $20 for Muse Wonder.

Sources

Every number on this page comes from these sources; models without a published number show none. Download the data. Cite as “Nexus Explorer, nexus.inc/models”.