BenchLeader

GLM 5 Turbo vs Muse Spark

Full profiles: GLM 5 Turbo (Zhipu AI) and Muse Spark (Meta): every benchmark, price and rank.

Updated · Data from 8 sources

As of 16 Nov 2026, there are not enough shared benchmarks for a verdict. Muse Spark leads in every category we measure both in.

Key facts (as of 16 Nov 2026)

  • GLM 5 Turbo costs $1.90 per 1M tokens blended (list price, 16 Nov 2026).
  • Of the 9 benchmarks both models have, Muse Spark leads 7 (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
Agents & tool useMuse Spark65 vs 63 category score; Terminal-Bench Hard 45.5% vs 33.3%Medium
Long documentsMuse Spark64 vs 60 category score; AA-LCR v1.1 78.0% vs 71.7%Medium
Following instructionsMuse Spark77 vs 72 category score; IFBench 75.9% vs 73.2%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

  • Not enough shared benchmarks for a verdict.
  • Muse Spark is stronger in agents & tools, instruction following, long context.

At a glance

SpecGLM 5 TurboMuse Spark
DeveloperZhipu AIMeta
Released15 Mar 20268 Apr 2026
Context window (tokens)200,000262,144
Compared atGLM 5 TurboMuse Spark
Open weightsNoNo
Price in / out / cached input per 1M$1.20 / $4.00 / $0.240–

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.

CategoryGLM 5 TurboMuse SparkLeader
Reasoning–67.4not comparable
Coding–65.1not comparable
Agents & tools62.865.0Muse Spark
Maths–59.1not comparable
Knowledge–63.1not comparable
Instruction following71.776.9Muse Spark
Human preference–68.1not comparable
Multimodal–64.2not comparable
Long context60.363.6Muse Spark

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 avg75Agents & toolsInstructionfollowingLong contextGLM 5 Turbo, Agents & tools: 62.8GLM 5 Turbo, Instruction following: 71.7GLM 5 Turbo, Long context: 60.3Muse Spark, Agents & tools: 65.0Muse Spark, Instruction following: 76.9Muse Spark, Long context: 63.6

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

Model, click to add or removeIndex

Where sources disagree

Across the 9 benchmarks both have, GLM 5 Turbo leads 2, Muse Spark leads 7.

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.

MeasureGLM 5 TurboMuse Spark
Blended price per 1M tokens$1.90–
Input / output per 1M$1.20 / $4.00– / –
Cached input per 1M$0.240–
Price per task: chat reply$0.0017–
Price per task: summarize a 30-page report$0.017–
Price per task: code edit$0.013–
Price per task: agentic coding session$0.088–
Price per task: structured extraction$0.0032–
Agentic coding session, input mostly cached$0.045–
Output speed26 tok/s–
Time to first answer4.2 s–

Each model is compared at its best-ranked reasoning-effort setting. GLM 5 Turbo and takes 4 s to its first answer token. Muse Spark scores 64.2. 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

55 benchmarks, grouped by category; each name links to the board and its leaders.

MetricPublisherGLM 5 TurboMuse Spark
BenchLeader IndexBenchLeader–64.2
Agents & tools scoreBenchLeader62.865.0
Instruction following scoreBenchLeader71.776.9
Long context scoreBenchLeader60.363.6
Coding scoreBenchLeader–65.1
Human preference scoreBenchLeader–68.1
Knowledge scoreBenchLeader–63.1
Maths scoreBenchLeader–59.1
Multimodal scoreBenchLeader–64.2
Reasoning scoreBenchLeader–67.4
Blended price $/MBenchLeader$1.90–
Output speedBenchLeader26 tok/s–
Time to first answerBenchLeader4.2 s–
Context windowBenchLeader200,000 tokens262,144 tokens
Reasoning
GPQA DiamondEpoch AI Benchmarking Hub–89.8%
Humanity's Last ExamScale AI / CAIS–40.6%
Arena Hard PromptsArena (LMArena)–1506
GPQA Diamond (AA)Artificial Analysis84.8%88.4%
Humanity's Last Exam (AA)Artificial Analysis27.8%40.7%
GPQA Diamond (Vals)Vals AI–89.7%
CritPtArtificial Analysis0.3%11.3%
Arena ExpertArena (LMArena)–1491
Coding
SciCodeSciCode–51.5%
Arena CodingArena (LMArena)–1529
SWE-bench (Vals)Vals AI–74.4%
SWE-Bench ProScale AI SEAL–55.0%
Vibe Code Bench v1.1Vals AI–19.7%
SWE-Bench Pro (private)Scale AI SEAL–44.7%
SWE Atlas v1.0: Codebase QnAScale AI SEAL–24.2%
SWE Atlas v1.0: Test WritingScale AI SEAL–31.1%
Terminal-Bench 2.1 (AA)Artificial Analysis–62.2%
Design Arena WebsiteDesign Arena12541253
Agents & tools
Terminal-Bench HardArtificial Analysis33.3%45.5%
τ²-Bench Telecom (AA)Artificial Analysis98.5%91.5%
MCP AtlasScale AI SEAL–82.2%
GDPval-AA v2.1Artificial Analysis–1006
Terminal-Bench 2.0 (Vals)Vals AI–59.5%
AA-Briefcase v1.1Artificial Analysis–646
Maths
OTIS Mock AIMEEpoch AI Benchmarking Hub–88.9%
Show all 55 benchmarks
MetricPublisherGLM 5 TurboMuse Spark
Maths
ProofBenchVals AI–17.0%
AIME (Vals)Vals AI–96.9%
Arena MathsArena (LMArena)–1469
Knowledge
MMLU-ProVals AI–87.3%
LegalBenchVals AI–84.2%
CorpFinVals AI–65.1%
TaxEvalVals AI–77.7%
PRBench FinanceScale AI SEAL–52.4%
PRBench LegalScale AI SEAL–52.3%
MultiNRCScale AI SEAL–59.0%
CaseLaw v2Vals AI–63.1%
MedCodeVals AI–51.3%
MedScribeVals AI–85.9%
Instruction following
IFBenchArtificial Analysis73.2%75.9%
MultiChallengeScale AI SEAL–75.5%
TutorBenchScale AI SEAL–68.5%
Arena Instruction FollowingArena (LMArena)–1465
Human preference
Arena TextArena (LMArena)–1489
Arena Creative WritingArena (LMArena)–1464
Arena Multi-turnArena (LMArena)–1493
Arena Longer QueriesArena (LMArena)–1476
Multimodal
Arena VisionArena (LMArena)–1306
MMMU-ProArtificial Analysis–80.5%
MMMU-Pro (Vals)Vals AI–87.4%
Long context
AA-LCR v1.1Artificial Analysis71.7%78.0%
Arena DocumentArena (LMArena)–1467
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub–153.1
AA Intelligence Index v4.3.2Artificial Analysis26.631.3
AA Progress Index v1Artificial Analysis277.3289.1
Safety & honesty
FORTRESSScale AI SEAL–20.2%

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

What changed

  • Muse Spark: 2 new results (GDPval-AA v2.1, AA-Briefcase v1.1).
  • Muse Spark: BenchLeader Index 64.6 → 64.1.
  • Muse Spark: 1 score revision.
  • GLM 5 Turbo: 1 new result (AA Progress Index v1).
  • Muse Spark: 1 new result (AA Progress Index v1).

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

GLM 5 Turbo vs Muse Spark: questions

Is GLM 5 Turbo or Muse Spark better for agents and tool use?
Muse Spark scores higher in agentic tasks (65 vs 63 on the category index, where 50 is average). On Terminal-Bench Hard, the top shared agents & tools benchmark, it is 33.3% for GLM 5 Turbo against 45.5% for Muse Spark.
Which is better for long documents?
Muse Spark scores higher on long-context benchmarks (64 vs 60). Muse Spark accepts more context: 262,144 against 200,000 tokens for GLM 5 Turbo.
Do independent benchmarks agree on GLM 5 Turbo vs Muse Spark?
Across the 9 benchmarks both have, GLM 5 Turbo leads 2, Muse Spark leads 7.
When was Muse Spark released?
Muse Spark was released on 8 Apr 2026.