BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases4

GPT-5.2 vs Muse Spark

Full profiles: GPT-5.2 (OpenAI) and Muse Spark (Meta): every benchmark, price and rank.

Updated · Data from 8 sources

As of 2 Nov 2026, GPT-5.2 and Muse Spark are neck and neck on quality (60.1 vs 64.3 on the BenchLeader Index, inside the ±5.2 margin). Choose GPT-5.2 for long context (66 vs 64); choose Muse Spark for agents & tools (66 vs 55), knowledge (63 vs 56), and reasoning (68 vs 61).

Key facts (as of 2 Nov 2026)

  • GPT-5.2 scores 60.1 on the BenchLeader Index and Muse Spark scores 64.3 (BenchLeader, 2 Nov 2026).
  • GPT-5.2 costs $4.81 per 1M tokens blended (list price, 2 Nov 2026).
  • Of the 24 benchmarks both models have, GPT-5.2 leads 12 (BenchLeader, 2 Nov 2026).
GPT-5.2 benchmarks xhigh OpenAI Muse Spark benchmarks MetaAdd 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 Spark65 vs 63 category score; SWE-bench (Vals) 74.4% vs 75.8%Medium
Agents & tool useMuse Spark66 vs 55 category score; MCP Atlas 82.2% vs 67.6%High
MathsMuse Spark59 vs 57 category score; ProofBench 17.0% vs 15.0%Medium
ReasoningMuse Spark68 vs 61 category score; GPQA Diamond 89.8% vs 91.4%Medium
KnowledgeMuse Spark63 vs 56 category score; MMLU-Pro 87.3% vs 86.2%Medium
Long documentsGPT-5.266 vs 64 category score; AA-LCR v1.1 82.7% vs 78.0%Medium
Following instructionsMuse Spark77 vs 74 category score; IFBench 75.9% vs 75.4%Medium
Chat quality (human preference)–Only Muse Spark has an Arena rating (68 category score)Not comparable
High volume on a budgetNeither flagshipA smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs–

Key differences

  • GPT-5.2 (xhigh) and Muse Spark are neck and neck on quality (60.1 vs 64.3, within the ±5.2 margin).
  • GPT-5.2 (xhigh) is stronger in long context.
  • Muse Spark is stronger in agents & tools, coding, instruction following, knowledge, maths, reasoning, human preference, multimodal.

At a glance

SpecGPT-5.2Muse Spark
DeveloperOpenAIMeta
Released11 Dec 20258 Apr 2026
Context window (tokens)400,000262,144
Compared atGPT-5.2 (xhigh)Muse Spark
Open weightsNoNo
Price in / out / cached input per 1M$1.75 / $14.00 / $0.175–

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.

CategoryGPT-5.2 (xhigh)Muse SparkLeader
Reasoning60.967.6Muse Spark
Coding62.964.9Muse Spark
Agents & tools54.766.1Muse Spark
Maths56.859.0Muse Spark
Knowledge56.062.9Muse Spark
Instruction following73.676.6Muse Spark
Human preference–68.1not comparable
Multimodal–64.2not comparable
Long context66.263.8GPT-5.2

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 & toolsMathsKnowledgeInstructionfollowingLong contextGPT-5.2, Reasoning: 60.9GPT-5.2, Coding: 62.9GPT-5.2, Agents & tools: 54.7GPT-5.2, Maths: 56.8GPT-5.2, Knowledge: 56.0GPT-5.2, Instruction following: 73.6GPT-5.2, Long context: 66.2Muse Spark, Reasoning: 67.6Muse Spark, Coding: 64.9Muse Spark, Agents & tools: 66.1Muse Spark, Maths: 59.0Muse Spark, Knowledge: 62.9Muse Spark, Instruction following: 76.6Muse Spark, Long context: 63.8

Not drawn, measured for fewer than two of these models: Multimodal, Human preference.

ModelIndex

Where sources disagree

Across the 24 benchmarks both have, GPT-5.2 leads 12, Muse Spark leads 11 and 1 is a tie. Every overall index that rates both puts Muse Spark ahead.

Overall indexGPT-5.2 (xhigh)Muse SparkLeader
BenchLeader Index60.164.3neck and neck
Arena Text1437†1489Muse Spark

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.

MeasureGPT-5.2 (xhigh)Muse Spark
Blended price per 1M tokens$4.81–
Input / output per 1M$1.75 / $14.00– / –
Cached input per 1M$0.175–
Price per task: chat reply$0.0049–
Price per task: summarize a 30-page report$0.029–
Price per task: code edit$0.032–
Price per task: agentic coding session$0.161–
Price per task: structured extraction$0.0063–
Agentic coding session, input mostly cached$0.090–
Output speed80 tok/s–
Time to first answer153.5 s–
Time to a complete answer159.7 s–

Each model is compared at its best-ranked reasoning-effort setting. GPT-5.2 (xhigh) scores 60.1 and takes 154 s to its first answer token; GPT-5.2 scores 57.9 in 2 s, GPT-5.2 (medium) scores 56.2 in 2 s. Muse Spark scores 64.3. 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

109 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.

MetricPublisherGPT-5.2 (xhigh)Muse Spark
BenchLeader IndexBenchLeader60.164.3
Agents & tools scoreBenchLeader54.766.1
Coding scoreBenchLeader62.964.9
Instruction following scoreBenchLeader73.676.6
Knowledge scoreBenchLeader56.062.9
Long context scoreBenchLeader66.263.8
Maths scoreBenchLeader56.859.0
Reasoning scoreBenchLeader60.967.6
Human preference scoreBenchLeader–68.1
Multimodal scoreBenchLeader–64.2
Blended price $/MBenchLeader$4.81–
Output speedBenchLeader80 tok/s–
Time to first answerBenchLeader153.5 s–
Context windowBenchLeader400,000 tokens262,144 tokens
Reasoning
GPQA DiamondEpoch AI Benchmarking Hub91.4%89.8%
Humanity's Last ExamScale AI / CAIS27.8%†40.6%
SimpleBenchSimpleBench45.8%†–
Arena Hard PromptsArena (LMArena)1497†1506
LiveBench 1.0 ReasoningLiveBench83.2%†–
GPQA Diamond (AA)Artificial Analysis90.3%88.4%
Humanity's Last Exam (AA)Artificial Analysis37.7%40.7%
GPQA Diamond (Vals)Vals AI91.7%89.7%
Kagi LLM BenchmarkKagi LLM Benchmark73.3%†–
ARC-AGI-1ARC Prize86.2%–
ARC-AGI-2ARC Prize52.9%–
CritPtArtificial Analysis11.6%11.3%
SciPredictScale AI SEAL20.6%†–
Chess PuzzlesEpoch AI Benchmarking Hub48.0%–
Mystery Game PuzzlesEpoch AI Benchmarking Hub22.0%†–
CL-benchEpoch AI Benchmarking Hub18.2%†–
LMCAEpoch AI Benchmarking Hub43.9%–
DTBenchEpoch AI Benchmarking Hub90.9%–
ForecastBenchForecastBench64.2%†–
Coding
SWE-bench Verified (Epoch)Epoch AI Benchmarking Hub73.8%†–
SciCodeSciCode–51.5%
WeirdMLWeirdML72.2%–
GSO-BenchGSO-Bench27.4%†–
Arena CodingArena (LMArena)1515†1529
Arena WebDevArena (LMArena)1416†–
Show all 109 benchmarks
MetricPublisherGPT-5.2 (xhigh)Muse Spark
Coding
LiveBench 1.0 CodingLiveBench76.1%†–
LiveCodeBenchVals AI85.4%–
SWE-bench (Vals)Vals AI75.8%74.4%
SWE-Bench ProScale AI SEAL29.9%†55.0%
SWE-bench Verified (bash only)SWE-bench69.0%†–
SWE-bench Verified (any scaffold)SWE-bench72.8%†–
Vibe Code Bench v1.1Vals AI53.5%19.7%
SWE-Bench Pro (private)Scale AI SEAL29.9%†55.0%
SWE Atlas: Codebase QnAScale AI SEAL–24.2%
SWE Atlas: Test WritingScale AI SEAL–31.1%
ALE-BenchEpoch AI Benchmarking Hub1249.8†–
AlgoTuneAlgoTune2.0†–
Terminal-Bench 2.1 (AA)Artificial Analysis–62.2%
LiveCodeBench (AA)Artificial Analysis88.9%–
Agents & tools
Terminal-BenchTerminal-Bench64.9%†–
GDPvalOpenAI49.7%†–
Remote Labor IndexScale AI / CAIS2.1%†–
LiveBench Agentic CodingLiveBench50.3%†–
Terminal-Bench HardArtificial Analysis47.0%45.5%
τ²-Bench Telecom (AA)Artificial Analysis84.8%91.5%
MCP AtlasScale AI SEAL67.6%82.2%
BFCL OverallBerkeley Function Calling Leaderboard55.9%†–
τ²-benchτ²-bench84.8%†–
τ³-bench Bankingτ²-bench32.2%†–
τ³-Banking (AA)Artificial Analysis11.1%†–
Poker AgentVals AI1131.8%†–
Terminal-Bench 2.0 (Vals)Vals AI51.7%†59.5%
DeepResearch BenchDeepResearch Bench41.1%†–
METR Time HorizonsMETR75.3%†–
Vending-Bench 2Andon Labs3591.3†–
Maths
FrontierMath Tiers 1–3Epoch AI Benchmarking Hub67.4%–
FrontierMath Tier 4Epoch AI Benchmarking Hub31.7%–
OTIS Mock AIMEEpoch AI Benchmarking Hub96.1%88.9%
ProofBenchVals AI15.0%17.0%
LiveBench MathematicsLiveBench93.2%†–
AIME (Vals)Vals AI96.9%96.9%
MGSMVals AI94.0%–
AIME 2026MathArena98.3%†–
HMMT February 2026MathArena97.0%†–
MathArena ApexMathArena13.5%†–
Arena MathsArena (LMArena)1454†1466
Knowledge
SimpleQA VerifiedEpoch AI Benchmarking Hub37.1%–
LiveBench 1.0 Data AnalysisLiveBench78.2%†–
MMLU-ProVals AI86.2%87.3%
LegalBenchVals AI82.8%84.2%
CorpFinVals AI65.9%65.1%
TaxEvalVals AI75.8%77.7%
MedQAVals AI94.1%–
PRBench FinanceScale AI SEAL–52.4%
PRBench LegalScale AI SEAL–52.3%
MultiNRCScale AI SEAL42.2%†59.0%
CaseLaw v2Vals AI66.0%63.1%
MedCodeVals AI49.8%51.3%
MedScribeVals AI84.4%85.9%
DrugDiscoveryBenchScale AI SEAL29.3%–
EBR-benchEpoch AI Benchmarking Hub23.0%–
Global-MMLU-Lite (AA)Artificial Analysis89.8%†–
Instruction following
LiveBench LanguageLiveBench79.8%†–
LiveBench Instruction FollowingLiveBench61.8%†–
IFBenchArtificial Analysis75.4%75.9%
MultiChallengeScale AI SEAL–75.5%
TutorBenchScale AI SEAL53.5%†68.5%
Arena Instruction FollowingArena (LMArena)1457†1464
Human preference
Arena TextArena (LMArena)1476†1489
Arena Creative WritingArena (LMArena)1433†1464
Arena Multi-turnArena (LMArena)1494†1493
Arena Longer QueriesArena (LMArena)1472†1476
Multimodal
Arena VisionArena (LMArena)1268†1306
MMMU-ProArtificial Analysis74.6%†80.5%
VISTAScale AI SEAL46.6%†–
MMMU-Pro (official)MMMU–80.4%†
MMMU-Pro (Vals)Vals AI86.7%87.4%
MortgageTaxVals AI67.1%–
VPCTEpoch AI Benchmarking Hub84.0%–
Long context
AA-LCR v1.1Artificial Analysis82.7%78.0%
Arena DocumentArena (LMArena)1425†1467
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub153.8†152.2
LiveBench 1.0LiveBench74.6%†–
AA Intelligence Index v4.3.2Artificial Analysis30.431.3
AA Progress Index v1Artificial Analysis295.5289.1
Vals RSI IndexVals AI15.8%–
Safety & honesty
FORTRESSScale AI SEAL17.5%†20.2%
MASKScale AI SEAL86.7%†–
PropensityBenchScale AI SEAL34.4%†–

Data as of 2026-11-02. Best configuration of each model; every score links to its source on the model pages.

What changed

  • GPT-5.2 (xhigh): 1 new result (AA Progress Index v1).
  • Muse Spark: 1 new result (AA Progress Index v1).
  • GPT-5.2 (xhigh): 1 new result (Vals RSI Index).
  • GPT-5.2 (xhigh): BenchLeader Index 60.8 → 60.3.
  • GPT-5.2 (xhigh): 1 score revision.
  • Muse Spark: 1 new result (PRBench Legal).
  • Muse Spark: 1 new result (SWE Atlas: Codebase QnA).
  • GPT-5.2 (xhigh): 1 new result (DrugDiscoveryBench).
  • Muse Spark: 2 new results (PRBench Legal, MultiNRC).
  • Muse Spark: BenchLeader Index 65.1 → 64.6.

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

GPT-5.2 vs Muse Spark: questions

Is GPT-5.2 better than Muse Spark?
GPT-5.2 (xhigh) and Muse Spark are neck and neck on quality (60.1 vs 64.3, within the ±5.2 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 2 Nov 2026, so check the category scores for your use.
Is GPT-5.2 or Muse Spark better for coding?
Muse Spark scores higher in coding (65 vs 63 on the category index, where 50 is average).
Is GPT-5.2 or Muse Spark better for agents and tool use?
Muse Spark scores higher in agentic tasks (66 vs 55 on the category index, where 50 is average). On MCP Atlas, the top shared agents & tools benchmark, it is 67.6% for GPT-5.2 against 82.2% for Muse Spark.
Which is better for long documents?
GPT-5.2 scores higher on long-context benchmarks (66 vs 64). GPT-5.2 accepts more context: 400,000 against 262,144 tokens for Muse Spark.
Do independent benchmarks agree on GPT-5.2 vs Muse Spark?
Across the 24 benchmarks both have, GPT-5.2 leads 12, Muse Spark leads 11 and 1 is a tie. Every overall index that rates both puts Muse Spark ahead.
More questions (1)
When was Muse Spark released?
Muse Spark was released on 8 Apr 2026.