Enter opens the highlighted result. Esc closes.

GPT-5.4 vs Muse Spark 1.2

GPT-5.4 leads Muse Spark 1.2 on coding, 64.4 to 55.7 on the Vibe Coding Index, and Muse Spark 1.2 costs 3.5x less per output token ($4.25 vs $15.00 per million). Updated October 22, 2026; 71 benchmarks have a result for both.

Replace GPT-5.4 Replace Muse Spark 1.2
GPT-5.4OpenAI
Coding
64.4
Overall
69.6
Output $/M
$15.00
22
Muse Spark 1.2Meta
Coding
55.7
Overall
60.4
Output $/M
$4.25
48
vs boards won of 71 shared, 1 tied

GPT-5.4 or Muse Spark 1.2: which to pick

Pick GPT-5.4 for coding: it scores 64.4 on the Vibe Coding Index against 55.7 for Muse Spark 1.2, though on 3 benchmarks against 5.

Pick Muse Spark 1.2 for output tokens 3.5x cheaper ($4.25 vs $15.00 per million) and a lower cost per solved task ($0.64 against $1.02).

Model AGPT-5.4OpenAI64.4/1003 independent benchmarks$1.02 per solved taskModel BMuse Spark 1.2Meta55.7/1005 independent benchmarks$0.64 per solved task
Shared results
71
GPT-5.4 only
39
Muse Spark 1.2 only
26
Like-for-like categories
8 of 9

Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
GPT-5.4 64.4 from 3 benchmarks, Muse Spark 1.2 55.7 from 5 benchmarks.
Widest category
Agentic work, 17.4 points apart on average over 5 shared benchmarks.
Widest single gap
Design Arena: 3D: GPT-5.4 1065 against Muse Spark 1.2 1301, 29.5 points apart on a 0 to 100 scale.
Output price
GPT-5.4 $15.00 against Muse Spark 1.2 $4.25 per million output tokens, 3.5x apart.
Context window
GPT-5.4 takes 1.05M tokens, Muse Spark 1.2 1.05M.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryGPT-5.4BarsMuse Spark 1.2Ties
Coding291 tie
Intelligence11
Agentic work14
Reasoning53
Math30
Knowledge29
Web design08
General814
All shared boards22481 tie
CodingLike-for-likeSix of these boards feed the Vibe Coding Index29
Coding benchmarks for GPT-5.4 and Muse Spark 1.2
BenchmarkGPT-5.4Muse Spark 1.2Leader
LiveBench codingNot measured78.2%Only Muse Spark 1.2 measured
LMArena WebDev14641534Muse Spark 1.2
AA Coding Index71.172.2Muse Spark 1.2
LiveCodeBench (Vals AI)84.1%Not measuredOnly GPT-5.4 measured
SWE-bench Verified (Vals AI)78.2%86.6%Muse Spark 1.2
Terminal-Bench 4 (Vals AI)Not measured6.1%Only Muse Spark 1.2 measured
Vibe Code Bench (Vals AI)67.4%79.1%Muse Spark 1.2
Code migration (Vals AI)35.0%29.9%GPT-5.4
ProgramBench (Vals AI)0.5%0.5%Tie
IOI (Vals AI)Not measured21.8%Only Muse Spark 1.2 measured
Terminal-Bench 2.1 (Vals AI)Not measured69.7%Only Muse Spark 1.2 measured
SkillsBench (Vals AI)51.7%53.0%Muse Spark 1.2
CyberBench (Vals AI)Not measured69.5%Only Muse Spark 1.2 measured
SciCodeNot measured57.4%Only Muse Spark 1.2 measured
Terminal-Bench Hard (Artificial Analysis run)57.6%Not measuredOnly GPT-5.4 measured
Terminal-Bench81.8%Not measuredOnly GPT-5.4 measured
FrontierSWENot measured12.0%Only Muse Spark 1.2 measured
GSO-Bench31.4%Not measuredOnly GPT-5.4 measured
WeirdML77.7%60.3%GPT-5.4
MirrorCode15.6%Not measuredOnly GPT-5.4 measured
ALE-Bench1607.0Not measuredOnly GPT-5.4 measured
LMArena Text: coding14961505Muse Spark 1.2
LMArena Text: software and IT14911505Muse Spark 1.2
LMArena WebDev: React14541526Muse Spark 1.2
LMArena WebDev: HTML14781543Muse Spark 1.2
LMArena WebDev: from an image1436Not measuredOnly GPT-5.4 measured
Neither model has a result on the other 13 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index11
Intelligence benchmarks for GPT-5.4 and Muse Spark 1.2
BenchmarkGPT-5.4Muse Spark 1.2Leader
AA Intelligence Index39.039.6Muse Spark 1.2
Epoch Capabilities Index157.3155.1GPT-5.4
Agentic workLike-for-likePublic boards, shown as published, kept out of the index14
Agentic work benchmarks for GPT-5.4 and Muse Spark 1.2
BenchmarkGPT-5.4Muse Spark 1.2Leader
AA Agentic IndexNot measured43.2Only Muse Spark 1.2 measured
Design Arena: agents building web apps10591221Muse Spark 1.2
Design Arena: agents building full-stack apps10171210Muse Spark 1.2
Design Arena: agents building mobile apps10871181Muse Spark 1.2
Design Arena: agents building Android apps10311222Muse Spark 1.2
Design Arena: agents building Godot games1135Not measuredOnly GPT-5.4 measured
Design Arena: agents building gamesNot measured1179Only Muse Spark 1.2 measured
Design Arena: HTML slidesNot measured1085Only Muse Spark 1.2 measured
Design Arena: PowerPoint slidesNot measured1172Only Muse Spark 1.2 measured
LiveBench researchNot measured67.7%Only Muse Spark 1.2 measured
LiveBench browserNot measured36.0%Only Muse Spark 1.2 measured
LiveBench long agenticNot measured21.3%Only Muse Spark 1.2 measured
SREBench (Vals AI)Not measured0.0%Only Muse Spark 1.2 measured
Finance Agent (Vals AI)Not measured60.6%Only Muse Spark 1.2 measured
Tax Agent Bench (Vals AI)Not measured32.8%Only Muse Spark 1.2 measured
Legal Research Bench (Vals AI)Not measured43.8%Only Muse Spark 1.2 measured
Excel Modeling Benchmark (Vals AI)Not measured57.0%Only Muse Spark 1.2 measured
BioMysteryBench (Vals AI)Not measured64.8%Only Muse Spark 1.2 measured
τ²-Bench Telecom87.1%Not measuredOnly GPT-5.4 measured
APEX-Agents52.4%45.4%GPT-5.4
METR time horizons74.3%Not measuredOnly GPT-5.4 measured
DeepResearch Bench35.1%Not measuredOnly GPT-5.4 measured
PostTrainBench19.0%Not measuredOnly GPT-5.4 measured
Vending-Bench 2$6,144Not measuredOnly GPT-5.4 measured
Neither model has a result on the other 17 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index53
Reasoning benchmarks for GPT-5.4 and Muse Spark 1.2
BenchmarkGPT-5.4Muse Spark 1.2Leader
GPQA Diamond91.7%Not measuredOnly GPT-5.4 measured
LiveBench reasoningNot measured90.0%Only Muse Spark 1.2 measured
MMMU Pro (Vals AI)87.5%86.1%GPT-5.4
GPQA Diamond (Artificial Analysis run)92.0%90.4%GPT-5.4
Humanity's Last Exam (Artificial Analysis run)43.7%45.5%Muse Spark 1.2
GPQA Diamond (Epoch AI run)93.3%Not measuredOnly GPT-5.4 measured
ARC-AGI93.7%Not measuredOnly GPT-5.4 measured
ARC-AGI-274.0%Not measuredOnly GPT-5.4 measured
SimpleBenchNot measured74.5%Only Muse Spark 1.2 measured
Chess puzzles48.0%27.0%GPT-5.4
Mystery game puzzles37.0%Not measuredOnly GPT-5.4 measured
EBR-bench25.4%Not measuredOnly GPT-5.4 measured
LMArena Text: hard prompts14851489Muse Spark 1.2
LMArena Text: expert15071475GPT-5.4
LMArena Text: hard prompts in English14791487Muse Spark 1.2
LMArena Vision: diagrams13271317GPT-5.4
LMArena Vision: homework1336Not measuredOnly GPT-5.4 measured
Kagi LLM Benchmark63.8%Not measuredOnly GPT-5.4 measured
Neither model has a result on the other 4 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index30
Math benchmarks for GPT-5.4 and Muse Spark 1.2
BenchmarkGPT-5.4Muse Spark 1.2Leader
AIME96.7%Not measuredOnly GPT-5.4 measured
OTIS Mock AIME97.8%Not measuredOnly GPT-5.4 measured
FrontierMath, tiers 1 to 378.6%Not measuredOnly GPT-5.4 measured
FrontierMath, tier 449.0%Not measuredOnly GPT-5.4 measured
ProofBench56.0%43.0%GPT-5.4
LMArena Text: math14881484GPT-5.4
LMArena Text: mathematical work14971480GPT-5.4
Neither model has a result on the other 6 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index29
Knowledge benchmarks for GPT-5.4 and Muse Spark 1.2
BenchmarkGPT-5.4Muse Spark 1.2Leader
MMLU Pro87.5%88.3%Muse Spark 1.2
LegalBench (Vals AI)86.0%85.3%GPT-5.4
CaseLaw (Vals AI)63.8%Not measuredOnly GPT-5.4 measured
CorpFin (Vals AI)65.3%70.9%Muse Spark 1.2
TaxEval (Vals AI)74.0%80.4%Muse Spark 1.2
MortgageTax (Vals AI)68.3%65.4%GPT-5.4
MedQA (Vals AI)96.1%Not measuredOnly GPT-5.4 measured
MedCode (Vals AI)41.3%49.4%Muse Spark 1.2
SimpleQA Verified45.1%60.3%Muse Spark 1.2
LMArena Text: science14791500Muse Spark 1.2
LMArena Text: medicine and healthcare14581496Muse Spark 1.2
LMArena Text: legal and government14861517Muse Spark 1.2
LMArena Text: business and finance14741484Muse Spark 1.2
Neither model has a result on the other 5 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index08
Web design benchmarks for GPT-5.4 and Muse Spark 1.2
BenchmarkGPT-5.4Muse Spark 1.2Leader
Design Arena: websites11981312Muse Spark 1.2
Design Arena: UI components12121298Muse Spark 1.2
Design Arena: game dev11941301Muse Spark 1.2
Design Arena: data viz11961334Muse Spark 1.2
Design Arena: 3D10651301Muse Spark 1.2
Design Arena: all code categories11791312Muse Spark 1.2
Design Arena: SVG11411284Muse Spark 1.2
Design Arena: ASCII art11881284Muse Spark 1.2
LiveBench designNot measured67.1%Only Muse Spark 1.2 measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index814
General benchmarks for GPT-5.4 and Muse Spark 1.2
BenchmarkGPT-5.4Muse Spark 1.2Leader
LiveBench data analysisNot measured76.5%Only Muse Spark 1.2 measured
Public Benefits Bench (Vals AI)Not measured68.5%Only Muse Spark 1.2 measured
SAGE (Vals AI)43.3%47.7%Muse Spark 1.2
MedScribe (Vals AI)77.5%90.1%Muse Spark 1.2
Vals Multimodal IndexNot measured69.8%Only Muse Spark 1.2 measured
IFBench73.9%Not measuredOnly GPT-5.4 measured
Long context reasoning (AA-LCR)82.0%79.0%GPT-5.4
DTBench94.4%94.7%Muse Spark 1.2
LMCA52.0%48.4%GPT-5.4
CL-bench27.9%Not measuredOnly GPT-5.4 measured
Furniture assembly37.5%Not measuredOnly GPT-5.4 measured
CL-bench Life21.7%Not measuredOnly GPT-5.4 measured
LMArena Text14691482Muse Spark 1.2
LMArena Text: instruction following14691462GPT-5.4
LMArena Text: creative writing14391436GPT-5.4
LMArena Vision13031304Muse Spark 1.2
LMArena Text: English14671483Muse Spark 1.2
LMArena Text: non-English14651476Muse Spark 1.2
LMArena Text: Chinese15201509GPT-5.4
LMArena Text: German1472Not measuredOnly GPT-5.4 measured
LMArena Text: Japanese1485Not measuredOnly GPT-5.4 measured
LMArena Text: Korean14491462Muse Spark 1.2
LMArena Text: French14931495Muse Spark 1.2
LMArena Text: Spanish14521484Muse Spark 1.2
LMArena Text: Russian14801483Muse Spark 1.2
LMArena Text: Polish1472Not measuredOnly GPT-5.4 measured
LMArena Text: multi-turn14821502Muse Spark 1.2
LMArena Text: longer queries14731473GPT-5.4
LMArena Text: writing and language14581453GPT-5.4
LMArena Text: entertainment and media14381443Muse Spark 1.2
LMArena Vision: English13011291GPT-5.4
LMArena Vision: Chinese1354Not measuredOnly GPT-5.4 measured
LMArena Vision: reading text13121313Muse Spark 1.2
LMArena Vision: humor1284Not measuredOnly GPT-5.4 measured
LMArena Vision: recognizing things1251Not measuredOnly GPT-5.4 measured
LMArena Vision: creative writing1283Not measuredOnly GPT-5.4 measured
LMArena Document1471Not measuredOnly GPT-5.4 measured
Neither model has a result on the other 9 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GPT-5.4 and Muse Spark 1.2
BenchmarkGPT-5.4Muse Spark 1.2Leader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

GPT-5.4$0.010Fits in one request
Muse Spark 1.2$0.0034Fits in one request

Muse Spark 1.2 costs less

Repository review

50K input + 3K output tokens

GPT-5.4$0.170Fits in one request
Muse Spark 1.2$0.075Fits in one request

Muse Spark 1.2 costs less

Agent loop

200K input + 10K output tokens

GPT-5.4$0.650Fits in one request
Muse Spark 1.2$0.292Fits in one request

Muse Spark 1.2 costs less

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of GPT-5.4 and Muse Spark 1.2
SpecificationOpenAI GPT-5.4Meta Muse Spark 1.2
MakerOpenAIMeta
WeightsNo public weights listedNo public weights listed
ReleasedMarch 5, 2026August 5, 2026
Context window1.05M1.05M
Input $/Mtok$2.50$1.25
Output $/Mtok$15.00$4.25
Coding index64.455.7
Independent measurements behind it35
Cost per solved task$1.02$0.64
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-5.4 better than Muse Spark 1.2 for coding?

Pick GPT-5.4 for coding: it scores 64.4 on the Vibe Coding Index against 55.7 for Muse Spark 1.2, though on 3 benchmarks against 5. Pick Muse Spark 1.2 for output tokens 3.5x cheaper ($4.25 vs $15.00 per million) and a lower cost per solved task ($0.64 against $1.02).

Which is cheaper, GPT-5.4 or Muse Spark 1.2?

GPT-5.4 costs $2.50 per million input tokens and $15.00 per million output tokens; Muse Spark 1.2 costs $1.25 and $4.25. Per solved SWE-bench Verified task, GPT-5.4 costs $1.02 and Muse Spark 1.2 $0.64.

Which has the bigger context window, GPT-5.4 or Muse Spark 1.2?

GPT-5.4, with 1.05M tokens against 1.05M.

How many benchmarks is the GPT-5.4 vs Muse Spark 1.2 comparison based on?

71 benchmarks have a result for both models as of October 22, 2026. GPT-5.4's coding index rests on 3 independent benchmarks and Muse Spark 1.2's on 5. Where only one model has a result, the page says so instead of naming a winner.

Compare next

Embed this comparison on your site

Free under CC BY 4.0, while the source link stays visible. Preview the frame · Terms and other embeds

<iframe src="https://llmbenchmark.ai/embed/compare/gpt-5-4-vs-meta-muse-spark-1-2" title="GPT-5.4 vs Muse Spark 1.2, LLMBenchmark.ai" width="100%" height="360" style="border:0;max-width:100%" loading="lazy"></iframe>
<p style="font:13px/1.4 sans-serif;margin:6px 0 0">Source: <a href="https://llmbenchmark.ai/compare/gpt-5-4-vs-meta-muse-spark-1-2">LLMBenchmark.ai, GPT-5.4 vs Muse Spark 1.2</a></p>