BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases7

Muse Spark 1.3 vs Step 5 Preview

Full profiles: Muse Spark 1.3 (Meta) and Step 5 Preview (StepFun): every benchmark, price and rank.

Updated · Data from 7 sources

As of 6 Nov 2026, Muse Spark 1.3 is ahead of Step 5 Preview on quality (62.0 vs 54.9 on the BenchLeader Index), but Step 5 Preview is 1.4× cheaper at $1.43 vs $2.00 per 1M tokens blended. Choose Muse Spark 1.3 for agents & tools (65 vs 51), coding (61 vs 50), and reasoning (61 vs 53); choose Step 5 Preview for long context (69 vs 66).

Key facts (as of 6 Nov 2026)

  • Muse Spark 1.3 scores 62.0 on the BenchLeader Index and Step 5 Preview scores 54.9 (BenchLeader, 6 Nov 2026).
  • Muse Spark 1.3 costs $2.00 per 1M tokens blended (list price, 6 Nov 2026).
  • Step 5 Preview costs $1.43 per 1M tokens blended (list price, 6 Nov 2026).
  • Of the 45 benchmarks both models have, Muse Spark 1.3 leads 39 (BenchLeader, 6 Nov 2026).
Muse Spark 1.3 benchmarks max Meta Step 5 Preview benchmarks StepFunAdd a third model

Which to pick for…

The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.

Use casePickWhyConfidence
CodingMuse Spark 1.361 vs 50 category score; FrontierCode 2.0 (AA) 23.1% vs 12.0%High
Agents & tool useMuse Spark 1.365 vs 51 category score; Arena Agent 2.5 vs -2.9High
MathsMuse Spark 1.363 vs 60 category score; Arena Maths 1508 vs 1466Medium
ReasoningMuse Spark 1.361 vs 53 category score; HLE-Diamond (AA) 33.7% vs 24.3%High
KnowledgeMuse Spark 1.359 vs 55 category score; AA-Omniscience v1.1 22.4 vs 16.4Medium
Long documentsStep 5 Preview69 vs 66 category score; AA-LCR v1.1 88.3% vs 83.0%Medium
Images and other mediaMuse Spark 1.366 vs 59 category score; Arena Vision 1308 vs 1265Medium
Chat quality (human preference)Muse Spark 1.369 vs 63 on Arena votesMedium
Lowest costStep 5 Preview$1.43 vs $2.00 per 1M blended; agentic coding session $0.071 vs $0.092–
Fastest first answerEither30.6 s (Muse Spark 1.3 (max)) vs 25.9 s (Step 5 Preview) to the first answer token–
High volume on a budgetNeither flagshipA smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs–

Key differences

  • Muse Spark 1.3 (max) leads on quality: 62.0 vs 54.9.
  • Muse Spark 1.3 (max) is stronger in agents & tools, coding, human preference, knowledge, maths, multimodal, reasoning.
  • Step 5 Preview is stronger in long context, composite.
  • Step 5 Preview is 1.4× cheaper ($1.43 vs $2.00 per 1M blended).

At a glance

SpecMuse Spark 1.3Step 5 Preview
DeveloperMetaStepFun
Released2 Sep 202618 Sep 2026
Context window (tokens)1,048,5761,000,000
Compared atMuse Spark 1.3 (max)Step 5 Preview
Open weightsNoYes
Price in / out / cached input per 1M$1.25 / $4.25 / $0.150$1.00 / $2.70 / $0.105

Category by category

Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.

CategoryMuse Spark 1.3 (max)Step 5 PreviewLeader
Reasoning60.852.7Muse Spark 1.3
Coding61.150.2Muse Spark 1.3
Agents & tools65.450.6Muse Spark 1.3
Maths63.159.7Muse Spark 1.3
Knowledge58.755.2Muse Spark 1.3
Human preference68.863.2Muse Spark 1.3
Multimodal65.759.4Muse Spark 1.3
Long context66.469.2Step 5 Preview
Composite–40.8not comparable

On one chart

Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.

2550 avg75ReasoningCodingAgents & toolsMathsKnowledgeLong contextMultimodalHumanpreferenceMuse Spark 1.3, Reasoning: 60.8Muse Spark 1.3, Coding: 61.1Muse Spark 1.3, Agents & tools: 65.4Muse Spark 1.3, Maths: 63.1Muse Spark 1.3, Knowledge: 58.7Muse Spark 1.3, Long context: 66.4Muse Spark 1.3, Multimodal: 65.7Muse Spark 1.3, Human preference: 68.8Step 5 Preview, Reasoning: 52.7Step 5 Preview, Coding: 50.2Step 5 Preview, Agents & tools: 50.6Step 5 Preview, Maths: 59.7Step 5 Preview, Knowledge: 55.2Step 5 Preview, Long context: 69.2Step 5 Preview, Multimodal: 59.4Step 5 Preview, Human preference: 63.2

Not drawn, measured for fewer than two of these models: Instruction following.

ModelIndex

Where sources disagree

Across the 45 benchmarks both have, Muse Spark 1.3 leads 39, Step 5 Preview leads 5 and 1 is a tie. Every overall index that rates both puts Muse Spark 1.3 ahead.

Overall indexMuse Spark 1.3 (max)Step 5 PreviewLeader
BenchLeader Index62.054.9Muse Spark 1.3
AA Intelligence Index v5.036.628.5Muse Spark 1.3
Arena Text14951448Muse Spark 1.3

Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).

Price, speed and latency

List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.

MeasureMuse Spark 1.3 (max)Step 5 Preview
Blended price per 1M tokens$2.00$1.43
Input / output per 1M$1.25 / $4.25$1.00 / $2.70
Cached input per 1M$0.150$0.105
Price per task: chat reply$0.0018$0.0012
Price per task: summarize a 30-page report$0.018$0.014
Price per task: code edit$0.014$0.010
Price per task: agentic coding session$0.092$0.071
Price per task: structured extraction$0.0034$0.0025
Agentic coding session, input mostly cached$0.043$0.030
Cost per benchmark task (AA Index)$2.39$2.35
Output tokens per task (AA Index)91k113k
Of which reasoning55k80k
Time per task (AA Index)13 min14 min
Output speed120 tok/s132 tok/s
Time to first answer30.6 s25.9 s
Time to a complete answer73.8 s57.3 s

Per token is not per task: an average task of Artificial Analysis' Intelligence Index cost $2.39 on Muse Spark 1.3 (max) and $2.35 on Step 5 Preview, and Step 5 Preview wrote 1.2× as many output tokens for the same work.

Each model is compared at its best-ranked reasoning-effort setting. Muse Spark 1.3 (max) scores 62.0 and takes 31 s to its first answer token; Muse Spark 1.3 (xhigh) scores 58.5 in 11 s, Muse Spark 1.3 scores 57.7 in 4 s. Step 5 Preview scores 54.9 and takes 26 s to its first answer token. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.

Every benchmark

117 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.

MetricPublisherMuse Spark 1.3 (max)Step 5 Preview
BenchLeader IndexBenchLeader62.054.9
Agents & tools scoreBenchLeader65.450.6
Coding scoreBenchLeader61.150.2
Human preference scoreBenchLeader68.863.2
Knowledge scoreBenchLeader58.755.2
Long context scoreBenchLeader66.469.2
Maths scoreBenchLeader63.159.7
Multimodal scoreBenchLeader65.759.4
Reasoning scoreBenchLeader60.852.7
Composite scoreBenchLeader–40.8
Blended price $/MBenchLeader$2.00$1.43
Output speedBenchLeader120 tok/s132 tok/s
Time to first answerBenchLeader30.6 s25.9 s
Context windowBenchLeader1,048,576 tokens1,000,000 tokens
Reasoning
Arena Hard PromptsArena (LMArena)15161477
LiveBench 2.0 ReasoningLiveBench89.7%†83.8%
LiveBench 1.0 ReasoningLiveBench89.7%†–
GPQA Diamond (AA)Artificial Analysis93.5%–
Humanity's Last Exam (AA)Artificial Analysis48.7%46.5%
HLE-Diamond (AA)Artificial Analysis33.7%24.3%
CritPtArtificial Analysis24.9%20.9%
CritPt v2 hard subset (AA)Artificial Analysis48.0%32.5%
MysteryMechanismVals AI36.0%18.5%
Chess PuzzlesEpoch AI Benchmarking Hub37.0%–
Mystery Game PuzzlesEpoch AI Benchmarking Hub25.0%–
LMCAEpoch AI Benchmarking Hub53.9%†–
DTBenchEpoch AI Benchmarking Hub96.5%†–
BTF-3Epoch AI Benchmarking Hub14.1%†–
Coding
SciCodeSciCode58.8%58.9%
Arena CodingArena (LMArena)15341506
Arena WebDevArena (LMArena)16561571†
LiveBench 2.0 CodingLiveBench90.2%†85.6%
LiveBench 1.0 CodingLiveBench81.1%†–
IOIVals AI56.6%38.1%
Code MigrationVals AI47.4%39.0%
Code Migration v1.1Vals AI–47.0%
ProgramBenchVals AI2.5%–
SWE-bench MultimodalVals AI43.1%40.8%
Vibe Code Bench 1-100Vals AI20.5%–
Show all 117 benchmarks
MetricPublisherMuse Spark 1.3 (max)Step 5 Preview
Coding
Vibe Code Bench v1.1Vals AI85.9%70.4%
CursorBenchCursor41.6%–
ALE-BenchEpoch AI Benchmarking Hub1688.3†–
FrontierSWEEpoch AI Benchmarking Hub25.8%–
Terminal-Bench 4.0 (AA)Artificial Analysis33.3%33.3%
Terminal-Bench 2.1 (AA)Artificial Analysis84.3%–
Startrise FrontendStartrise74.7%†–
GameBench 2 OverallDuelLab GameBench 257.8†–
GameBench 2 by reasoning settingDuelLab GameBench 268.7†–
FrontierCode 1.1 (AA)Artificial Analysis40.6%38.4%
FrontierCode 2.0 (AA)Artificial Analysis23.1%12.0%
Agents & tools
Terminal-BenchTerminal-Bench14.6%†–
APEX-AgentsMercor47.6%–
Arena AgentArena (LMArena)2.5-2.9
LiveBench Long AgenticLiveBench25.5%†19.4%
LiveBench BrowserLiveBench30.9%†24.8%
LiveBench Agentic CodingLiveBench64.1%†–
Terminal-Bench 2.1 (Vals)Vals AI79.0%–
HiL-BenchScale AI SEAL66.9%†–
GDPval-AA v2.1, normalised scoreArtificial Analysis–54.3%
GDPval-AA v2.1Artificial Analysis–1594
τ³-Banking (AA)Artificial Analysis50.5%–
ITBench SRE (AA)Artificial Analysis33.2%55.6%
Analyst Agent (AA)Artificial Analysis37.5%35.0%
APEX-Agents (AA)Artificial Analysis–38.0%
CUA-benchVals AI5.8%–
CyberBenchVals AI72.7%–
Finance Agent v2Vals AI60.0%50.7%
Finance Agent v2.1Vals AI–58.9%
Harvey's Legal Agent Benchmark v1.1Vals AI–12.9%
Harvey's Legal Agent Benchmark v1Vals AI23.8%9.2%
Legal Research BenchVals AI55.3%40.4%
SREBenchVals AI3.8%0.4%
Tax Agent BenchVals AI42.4%27.3%
Terminal-Bench 4.0 (Vals)Vals AI24.8%31.8%
Terminal-Bench ScienceVals AI10.0%2.9%
AutomationBenchArtificial Analysis57.9%51.0%
GDP.pdf (AA)Artificial Analysis26.6%14.8%
Harvey LAB-AA v1.1Harvey8.9%–
EnterpriseOps-GymArtificial Analysis–47.2%
AA-Briefcase v1.1Artificial Analysis15811425
Terminal-Bench-Science 0.1 (AA)Artificial Analysis10.9%2.4%
CWE-Bench (AA)Artificial Analysis–52.5%
DeepSec-Bench (AA)Artificial Analysis–15.8%
CyberGym end-to-end (AA)Artificial Analysis–74.8%
Maths
FrontierMath Tiers 1–3Epoch AI Benchmarking Hub74.0%–
FrontierMath Tier 4Epoch AI Benchmarking Hub46.3%–
OTIS Mock AIMEEpoch AI Benchmarking Hub99.2%†–
ProofBenchVals AI58.0%42.0%
LiveBench MathematicsLiveBench96.0%†–
Arena MathsArena (LMArena)15081466
Knowledge
LiveBench 2.0 Data AnalysisLiveBench79.6%†63.5%
LiveBench ResearchLiveBench68.6%†72.8%
LiveBench 1.0 Data AnalysisLiveBench79.6%†–
AA-Omniscience v1.1Artificial Analysis22.416.4
PRBench FinanceScale AI SEAL59.5%†–
PRBench LegalScale AI SEAL61.6%†–
BioMysteryBenchVals AI–67.4%
Excel Modeling BenchmarkVals AI67.4%57.1%
Excel Modeling Benchmark v1.1Vals AI–69.5%
Public Benefits Bench v1.1Vals AI–58.1%
DrugDiscoveryBenchScale AI SEAL62.2%†–
GDP.pdfEpoch AI Benchmarking Hub27.6%†–
AA-Omniscience v1.1: accuracyArtificial Analysis38.0%36.4%
AA-Omniscience v1.1: non-hallucinationArtificial Analysis74.8%68.7%
Instruction following
LiveBench LanguageLiveBench82.8%†–
LiveBench Instruction FollowingLiveBench78.0%†–
Arena Instruction FollowingArena (LMArena)14831446
Human preference
Arena TextArena (LMArena)14951448
Arena Creative WritingArena (LMArena)14641411
Arena Multi-turnArena (LMArena)14931442
Arena Longer QueriesArena (LMArena)14961460
Multimodal
Arena VisionArena (LMArena)13081265
LiveBench DesignLiveBench74.8%†76.9%
MMMU-ProArtificial Analysis82.0%†76.4%
Code2Video-AAArtificial Analysis1019821
AA-3DArtificial Analysis1033854
Long context
AA-LCR v1.1Artificial Analysis83.0%88.3%
Arena DocumentArena (LMArena)1468–
MLCRArtificial Analysis13.9%18.3%
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub158.0†–
LiveBench 2.0LiveBench63.0%†58.6%
LiveBench 1.0LiveBench81.6%†–
AA Intelligence Index v5.0Artificial Analysis36.628.5
AA Intelligence Index v4.3.2Artificial Analysis48.143.7
AA Progress Index v1Artificial Analysis324.5318.2
Vals IndexVals AI–54.8
Vals Finance IndexVals AI–55.4%
Vals RSI IndexVals AI19.6%–
Vals Science IndexVals AI–29.6%
AA Cyber IndexArtificial Analysis–47.7
Metadata
AA Openness IndexArtificial Analysis–33.3

Data as of 2026-11-06. Best configuration of each model; every score links to its source on the model pages.

DuelLab GameBench 2, release gb2-20261103-09b2218cbd85 (https://benchmarks.duellab.org/releases/gb2-20261103-09b2218cbd85/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.

What changed

  • Muse Spark 1.3 (max): 1 score revision.
  • Step 5 Preview: 2 new results (Vals Index, Harvey's Legal Agent Benchmark v1.1).
  • Muse Spark 1.3 (max): 1 score revision.
  • Step 5 Preview: 1 score revision.
  • Muse Spark 1.3 (max): BenchLeader Index 64.5 → 61.8.
  • Step 5 Preview: BenchLeader Index 58.0 → 54.7.
  • Muse Spark 1.3 (max): 5 new results (AA Intelligence Index v5.0, AA Progress Index v1, HLE-Diamond (AA)…); 3 score revisions; cost per task moved from $1.31/M to $2.39/M (81% up).
  • Step 5 Preview: 5 new results (AA Intelligence Index v5.0, AA Progress Index v1, HLE-Diamond (AA)…); 2 score revisions; Step 5 Preview cost per task moved from $1.03/M to $2.35/M (128% up).
  • Muse Spark 1.3 (max): cost per task moved from $3.69/M to $1.31/M (64% down).
  • Muse Spark 1.3 (max): cost per task moved from $1.60/M to $3.69/M (130% up).

Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs

Muse Spark 1.3 vs Step 5 Preview: questions

Is Muse Spark 1.3 better than Step 5 Preview?
Muse Spark 1.3 (max) leads on quality: 62.0 vs 54.9. The BenchLeader Index combines every independent quality benchmark; Muse Spark 1.3 (max) is ahead overall as of 6 Nov 2026, but check the category scores for your use.
Muse Spark 1.3 vs Step 5 Preview: which is cheaper, and will that change?
Step 5 Preview is cheaper today: $1.43 against $2.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
Is Muse Spark 1.3 or Step 5 Preview better for coding?
Muse Spark 1.3 scores higher in coding (61 vs 50 on the category index, where 50 is average). On FrontierCode 2.0 (AA), the top shared coding benchmark, it is 23.1% for Muse Spark 1.3 against 12.0% for Step 5 Preview.
Which is faster, Muse Spark 1.3 or Step 5 Preview?
Step 5 Preview streams faster (132 vs 120 output tokens per second, Artificial Analysis 7-day median).
Is Muse Spark 1.3 or Step 5 Preview better for agents and tool use?
Muse Spark 1.3 scores higher in agentic tasks (65 vs 51 on the category index, where 50 is average). On Arena Agent, the top shared agents & tools benchmark, it is 2.5 for Muse Spark 1.3 against -2.9 for Step 5 Preview.
More questions (4)
Which is better for long documents?
Step 5 Preview scores higher on long-context benchmarks (69 vs 66). Both take about 1M tokens: Muse Spark 1.3 1,048,576, Step 5 Preview 1,000,000.
Which has lower latency?
Step 5 Preview answers first: 25.9 s against 30.6 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
Do independent benchmarks agree on Muse Spark 1.3 vs Step 5 Preview?
Across the 45 benchmarks both have, Muse Spark 1.3 leads 39, Step 5 Preview leads 5 and 1 is a tie. Every overall index that rates both puts Muse Spark 1.3 ahead.
When was Step 5 Preview released?
Step 5 Preview was released on 18 Sep 2026.