BenchLeader

Muse Spark Open vs Nova 2.0 Pro Preview

Full profiles: Muse Spark Open (Meta) and Nova 2.0 Pro Preview (Amazon): every benchmark, price and rank.

Updated · Data from 10 sources

As of 16 Nov 2026, Muse Spark Open and Nova 2.0 Pro Preview are neck and neck on quality (53.5 vs 53.4 on the BenchLeader Index, inside the ±12.2 margin), and Muse Spark Open is 1.7× cheaper at $2.00 vs $3.44 per 1M tokens blended. Choose Muse Spark Open for multimodal (63 vs 47) and long context (64 vs 56) and fast first answers (9.9 s vs 16.1 s), where it also streams ~1.5× faster; choose Nova 2.0 Pro Preview for coding (43 vs 36).

Key facts (as of 16 Nov 2026)

  • Muse Spark Open scores 53.5 on the BenchLeader Index and Nova 2.0 Pro Preview scores 53.4 (BenchLeader, 16 Nov 2026).
  • Muse Spark Open costs $2.00 per 1M tokens blended (list price, 16 Nov 2026).
  • Nova 2.0 Pro Preview costs $3.44 per 1M tokens blended (list price, 16 Nov 2026).
  • Muse Spark Open streams ~1.5× faster than Nova 2.0 Pro Preview (Artificial Analysis, 7-day median, 16 Nov 2026).
  • Of the 9 benchmarks both models have, Muse Spark Open leads 9 (BenchLeader, 16 Nov 2026).

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
CodingNova 2.0 Pro Preview43 vs 36 category score; Terminal-Bench 2.1 (AA) 29.6% vs 76.4%Medium
Long documentsMuse Spark Open64 vs 56 category score; AA-LCR v1.1 78.3% vs 64.0%High
Images and other mediaMuse Spark Open63 vs 47 category score; MMMU-Pro 80.9% vs 64.5%High
Lowest costMuse Spark Open$2.00 vs $3.44 per 1M blended; agentic coding session $0.092 vs $0.115–
Fastest first answerMuse Spark Open9.9 s (Muse Spark Open (xhigh)) vs 16.1 s (Nova 2.0 Pro Preview (medium)) 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 Open (xhigh) and Nova 2.0 Pro Preview (medium) are neck and neck on quality (53.5 vs 53.4, within the ±12.2 margin).
  • Muse Spark Open (xhigh) is stronger in long context, multimodal.
  • Nova 2.0 Pro Preview (medium) is stronger in coding.
  • Muse Spark Open (xhigh) is 1.7× cheaper ($2.00 vs $3.44 per 1M blended).
  • Muse Spark Open (xhigh) streams 1.8× faster (180 vs 98 tokens per second).

At a glance

SpecMuse Spark OpenNova 2.0 Pro Preview
DeveloperMetaAmazon
Released15 Oct 20262 Dec 2025
Context window (tokens)1,048,576256,000
Compared atMuse Spark Open (xhigh)Nova 2.0 Pro Preview (medium)
Open weightsYesNo
Price in / out / cached input per 1M$1.25 / $4.25 / $0.150$1.25 / $10.00 / –

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 Open (xhigh)Nova 2.0 Pro Preview (medium)Leader
Reasoning51.3–not comparable
Coding36.043.1Nova 2.0 Pro Preview
Agents & tools–54.8not comparable
Knowledge58.8–not comparable
Instruction following–76.8not comparable
Multimodal63.346.5Muse Spark Open
Long context63.856.3Muse Spark Open

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 avg75CodingLong contextMultimodalMuse Spark Open, Coding: 36.0Muse Spark Open, Long context: 63.8Muse Spark Open, Multimodal: 63.3Nova 2.0 Pro Preview, Coding: 43.1Nova 2.0 Pro Preview, Long context: 56.3Nova 2.0 Pro Preview, Multimodal: 46.5

Not drawn, measured for fewer than two of these models: Reasoning, Agents & tools, Maths, Knowledge, Instruction following, Human preference.

Model, click to add or removeIndex

Where sources disagree

Across the 9 benchmarks both have, Muse Spark Open leads 9, Nova 2.0 Pro Preview leads 0.

Overall indexMuse Spark Open (xhigh)Nova 2.0 Pro Preview (medium)Leader
BenchLeader Index53.553.4neck and neck

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 Open (xhigh)Nova 2.0 Pro Preview (medium)
Blended price per 1M tokens$2.00$3.44
Input / output per 1M$1.25 / $4.25$1.25 / $10.00
Cached input per 1M$0.150–
Price per task: chat reply$0.0018$0.0035
Price per task: summarize a 30-page report$0.018$0.021
Price per task: code edit$0.014$0.022
Price per task: agentic coding session$0.092$0.115
Price per task: structured extraction$0.0034$0.0045
Agentic coding session, input mostly cached$0.043–
Cost per benchmark task (AA Index)$2.02–
Output tokens per task (AA Index)66k–
Of which reasoning32k–
Time per task (AA Index)6 min–
Output speed180 tok/s98 tok/s
Time to first answer9.9 s16.1 s
Time to a complete answer, last measured 13 Nov 202639.9 s32.3 s

Each model is compared at its best-ranked reasoning-effort setting. Muse Spark Open (xhigh) scores 53.5 and takes 10 s to its first answer token. Nova 2.0 Pro Preview (medium) scores 53.4 and takes 16 s to its first answer token; Nova 2.0 Pro Preview (low) scores 51.9 in 7 s. 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

82 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 Open (xhigh)Nova 2.0 Pro Preview (medium)
BenchLeader IndexBenchLeader53.553.4
Coding scoreBenchLeader36.043.1
Knowledge scoreBenchLeader58.8–
Long context scoreBenchLeader63.856.3
Multimodal scoreBenchLeader63.346.5
Reasoning scoreBenchLeader51.3–
Agents & tools scoreBenchLeader–54.8
Instruction following scoreBenchLeader–76.8
Blended price $/MBenchLeader$2.00$3.44
Output speedBenchLeader180 tok/s98 tok/s
Time to first answerBenchLeader9.9 s16.1 s
Context windowBenchLeader1,048,576 tokens256,000 tokens
Reasoning
LiveBench 2.0 ReasoningLiveBench73.7%†–
GPQA Diamond (AA)Artificial Analysis89.0%78.5%
Humanity's Last Exam (AA)Artificial Analysis43.9%9.4%
HLE-Diamond (AA)Artificial Analysis26.3%–
ARC-AGI-1ARC Prize89.0%–
ARC-AGI-2ARC Prize57.5%–
ARC-AGI-3ARC Prize0.4%–
CritPtArtificial Analysis14.9%0.0%
CritPt v2 hard subset (AA)Artificial Analysis29.5%–
HLE-Diamond (Scale)Scale AI SEAL28.0%†–
Coding
SciCodeSciCode–42.7%
Arena WebDevArena (LMArena)1526†–
LiveBench 2.0 CodingLiveBench55.1%†–
SciCode (AA)Artificial Analysis58.5%–
Code MigrationVals AI21.2%–
Code Migration v1.1Vals AI26.9%–
Vibe Code Bench v1.1Vals AI76.2%–
SWE Atlas v1.1Scale AI SEAL42.3%†–
SWE Atlas v1.1: Codebase QnAScale AI SEAL50.0%†–
SWE Atlas v1.1: Test WritingScale AI SEAL41.1%†–
SWE Atlas v1.1: RefactoringScale AI SEAL35.7%†–
SWE-Bench Pro V2Scale AI SEAL95.0%–
SWE-Bench Pro V2 HARDScale AI SEAL82.3%–
Terminal-Bench 4.0 (AA)Artificial Analysis8.1%–
Terminal-Bench 2.1 (AA)Artificial Analysis76.4%29.6%
Show all 82 benchmarks
MetricPublisherMuse Spark Open (xhigh)Nova 2.0 Pro Preview (medium)
Coding
GameBench 2 OverallDuelLab GameBench 255.3†–
GameBench 2 by reasoning settingDuelLab GameBench 259.3–
Design Arena WebsiteDesign Arena1305–
FrontierCode 1.1 (AA)Artificial Analysis27.6%–
FrontierCode 2.0 (AA)Artificial Analysis7.6%–
LiveCodeBench (AA)Artificial Analysis–73.0%
Agents & tools
LiveBench Long AgenticLiveBench12.1%†–
LiveBench BrowserLiveBench26.2%†–
Terminal-Bench HardArtificial Analysis–24.2%
τ²-Bench Telecom (AA)Artificial Analysis–92.7%
HiL-BenchScale AI SEAL59.7%†–
GDPval-AA v2.1, normalised scoreArtificial Analysis45.1%0.0%
GDPval-AA v2.1Artificial Analysis1401477
Finance Agent v2Vals AI61.3%–
Finance Agent v2.1Vals AI60.7%–
Harvey's Legal Agent Benchmark v1.1Vals AI25.0%–
Harvey's Legal Agent Benchmark v1Vals AI16.7%–
Legal Research BenchVals AI38.5%–
Tax Agent BenchVals AI26.3%–
Terminal-Bench 4.0 (Vals)Vals AI6.6%–
AutomationBenchArtificial Analysis49.7%–
GDP.pdf (AA)Artificial Analysis18.6%–
AA-Briefcase v1.1Artificial Analysis1280–
GDPxGDPx by SemiAnalysis1.7%†–
Terminal-Bench-Science 0.1 (AA)Artificial Analysis1.9%–
Maths
ArXivMathMathArena27.0%–
BrokenArXivMathArena30.1%–
AIME 2025 (AA)Artificial Analysis–89.0%
Knowledge
LiveBench 2.0 Data AnalysisLiveBench60.2%†–
LiveBench ResearchLiveBench82.7%†–
AA-Omniscience v1.1Artificial Analysis25.8–
PRBench LegalScale AI SEAL59.8%†–
Excel Modeling BenchmarkVals AI53.7%–
Excel Modeling Benchmark v1.1Vals AI62.0%–
GDP.pdfEpoch AI Benchmarking Hub–2.0%†
AA-Omniscience v1.1: accuracyArtificial Analysis42.0%–
AA-Omniscience v1.1: non-hallucinationArtificial Analysis72.0%–
Global-MMLU-Lite (AA)Artificial Analysis–87.6%†
Instruction following
IFBenchArtificial Analysis–79.0%
MultiChallengeScale AI SEAL74.2%†–
Multimodal
LiveBench DesignLiveBench58.1%†–
MMMU-ProArtificial Analysis80.9%64.5%
SAGEVals AI85.8%–
Humanity's Sixth SenseScale AI SEAL29.6%†–
Long context
AA-LCR v1.1Artificial Analysis78.3%64.0%
MLCRArtificial Analysis7.8%–
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub155.4†–
LiveBench 2.0LiveBench48.5%†–
AA Intelligence Index v5.0.1Artificial Analysis24.9–
AA Intelligence Index v4.3.2Artificial Analysis38.914.2
AA Progress Index v1Artificial Analysis306.7244.3
Vals IndexVals AI48.6–
Vals Finance IndexVals AI53.0%–
Vals Legal IndexVals AI68.5%–
Safety & honesty
FORTRESSScale AI SEAL7.6%†–
DistressBenchScale AI SEAL87.7%†–
Metadata
AA Openness IndexArtificial Analysis27.8–

Data as of 2026-11-16. Best configuration of each model; every score links to its source on the model pages.

DuelLab GameBench 2, release gb2-20261111-7b598bb13de4 (https://benchmarks.duellab.org/releases/gb2-20261111-7b598bb13de4/), 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 Open (xhigh): 1 score revision.
  • Muse Spark Open (xhigh): 1 new result (GameBench 2 by reasoning setting); 1 score revision.
  • Muse Spark Open (xhigh): BenchLeader Index 53.3 → 53.8.
  • Muse Spark Open (xhigh): 1 new result (ARC-AGI-3).
  • Muse Spark Open (xhigh): BenchLeader Index 59.5 → 53.3.
  • Nova 2.0 Pro Preview (medium): BenchLeader Index 52.4 → 53.6.
  • Muse Spark Open (xhigh): 6 new results (AA Intelligence Index v5.0.1, AA Progress Index v1, HLE-Diamond (AA)…); 3 score revisions; cost per task moved from $0.84/M to $2.02/M (139% up).
  • Nova 2.0 Pro Preview (medium): 1 new result (AA Progress Index v1).
  • Muse Spark Open (xhigh): 6 new results (MLCR, Code Migration v1.1, Excel Modeling Benchmark v1.1…); cached input price moved from $1.25/M to $0.15/M (88% down).
  • Muse Spark Open (xhigh): 6 new results (FrontierCode 1.1 (AA), Terminal-Bench-Science 0.1 (AA), Vals Index…).

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

Muse Spark Open vs Nova 2.0 Pro Preview: questions

Is Muse Spark Open better than Nova 2.0 Pro Preview?
Muse Spark Open (xhigh) and Nova 2.0 Pro Preview (medium) are neck and neck on quality (53.5 vs 53.4, within the ±12.2 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 16 Nov 2026, so check the category scores for your use.
Muse Spark Open vs Nova 2.0 Pro Preview: which is cheaper, and will that change?
Muse Spark Open is cheaper today: $2.00 against $3.44 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
Is Muse Spark Open or Nova 2.0 Pro Preview better for coding?
Nova 2.0 Pro Preview scores higher in coding (43 vs 36 on the category index, where 50 is average).
Which is faster, Muse Spark Open or Nova 2.0 Pro Preview?
Muse Spark Open streams ~1.5× faster (180 vs 98 output tokens per second, Artificial Analysis 7-day median).
Which is better for long documents?
Muse Spark Open scores higher on long-context benchmarks (64 vs 56). Muse Spark Open accepts more context: 1,048,576 against 256,000 tokens for Nova 2.0 Pro Preview.
More questions (3)
Which has lower latency?
Muse Spark Open answers first: 9.9 s against 16.1 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 Open vs Nova 2.0 Pro Preview?
Across the 9 benchmarks both have, Muse Spark Open leads 9, Nova 2.0 Pro Preview leads 0.
When was Muse Spark Open released?
Muse Spark Open was released on 15 Oct 2026.