Enter opens the highlighted result. Esc closes.

GPT-5.1 vs GPT-5.4

GPT-5.4 leads GPT-5.1 on coding, 64.4 to 40.4 on the Vibe Coding Index, and GPT-5.1 costs 1.5x less per output token ($10.00 vs $15.00 per million). Updated October 22, 2026; 91 benchmarks have a result for both.

Replace GPT-5.1 Replace GPT-5.4
GPT-5.1OpenAI
Coding
40.4
Overall
55.0
Output $/M
$10.00
10
GPT-5.4OpenAI
Coding
64.4
Overall
69.6
Output $/M
$15.00
81
vs boards won of 91 shared

GPT-5.1 or GPT-5.4: which to pick

Pick GPT-5.4 for coding: it scores 64.4 on the Vibe Coding Index against 40.4 for GPT-5.1.

Pick GPT-5.1 for output tokens 1.5x cheaper ($10.00 vs $15.00 per million) and work measured by LiveCodeBench (Vals AI) or DeepResearch Bench, where it leads.

Model AGPT-5.1OpenAI40.4/1003 independent benchmarks$1.02 per solved taskModel BGPT-5.4OpenAI64.4/1003 independent benchmarks$1.02 per solved task
Shared results
91
GPT-5.1 only
10
GPT-5.4 only
19
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-5.1 40.4 from 3 benchmarks, GPT-5.4 64.4 from 3 benchmarks.
Widest category
Coding, 15.6 points apart on average over 14 shared benchmarks.
Widest single gap
ARC-AGI-2: GPT-5.1 17.6% against GPT-5.4 74.0%, 56.3 points apart on a 0 to 100 scale.
Output price
GPT-5.1 $10.00 against GPT-5.4 $15.00 per million output tokens, 1.5x apart.
Context window
GPT-5.1 takes 400K tokens, GPT-5.4 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-5.1BarsGPT-5.4Ties
Coding113
Intelligence02
Agentic work12
Reasoning014
Math04
Knowledge58
Web design26
General132
All shared boards1081
CodingLike-for-likeSix of these boards feed the Vibe Coding Index113
Coding benchmarks for GPT-5.1 and GPT-5.4
BenchmarkGPT-5.1GPT-5.4Leader
SWE-bench Verified66.0%Not measuredOnly GPT-5.1 measured
LMArena WebDev13961464GPT-5.4
AA Coding Index49.471.1GPT-5.4
LiveCodeBench (Vals AI)86.5%84.1%GPT-5.1
SWE-bench Verified (Vals AI)69.8%78.2%GPT-5.4
Vibe Code Bench (Vals AI)24.6%67.4%GPT-5.4
Code migration (Vals AI)Not measured35.0%Only GPT-5.4 measured
ProgramBench (Vals AI)Not measured0.5%Only GPT-5.4 measured
SkillsBench (Vals AI)Not measured51.7%Only GPT-5.4 measured
LiveCodeBench (Artificial Analysis run)86.8%Not measuredOnly GPT-5.1 measured
Terminal-Bench Hard (Artificial Analysis run)45.5%57.6%GPT-5.4
Terminal-Bench47.6%81.8%GPT-5.4
GSO-Bench13.7%31.4%GPT-5.4
WeirdML60.8%77.7%GPT-5.4
MirrorCodeNot measured15.6%Only GPT-5.4 measured
ALE-BenchNot measured1607.0Only GPT-5.4 measured
LMArena Text: coding14531496GPT-5.4
LMArena Text: software and IT14541491GPT-5.4
LMArena WebDev: React13071454GPT-5.4
LMArena WebDev: HTML14051478GPT-5.4
LMArena WebDev: from an image14211436GPT-5.4
Neither model has a result on the other 18 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for GPT-5.1 and GPT-5.4
BenchmarkGPT-5.1GPT-5.4Leader
AA Intelligence Index24.739.0GPT-5.4
Epoch Capabilities Index149.8157.3GPT-5.4
Agentic workLike-for-likePublic boards, shown as published, kept out of the index12
Agentic work benchmarks for GPT-5.1 and GPT-5.4
BenchmarkGPT-5.1GPT-5.4Leader
Design Arena: agents building web appsNot measured1059Only GPT-5.4 measured
Design Arena: agents building full-stack appsNot measured1017Only GPT-5.4 measured
Design Arena: agents building mobile appsNot measured1087Only GPT-5.4 measured
Design Arena: agents building Android appsNot measured1031Only GPT-5.4 measured
Design Arena: agents building Godot gamesNot measured1135Only GPT-5.4 measured
τ²-Bench Telecom81.9%87.1%GPT-5.4
APEX-AgentsNot measured52.4%Only GPT-5.4 measured
METR time horizonsNot measured74.3%Only GPT-5.4 measured
DeepResearch Bench42.8%35.1%GPT-5.1
PostTrainBenchNot measured19.0%Only GPT-5.4 measured
Vending-Bench 2$1,473$6,144GPT-5.4
Neither model has a result on the other 30 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index014
Reasoning benchmarks for GPT-5.1 and GPT-5.4
BenchmarkGPT-5.1GPT-5.4Leader
GPQA Diamond86.6%91.7%GPT-5.4
MMMU Pro (Vals AI)83.2%87.5%GPT-5.4
GPQA Diamond (Artificial Analysis run)87.3%92.0%GPT-5.4
Humanity's Last Exam (Artificial Analysis run)28.5%43.7%GPT-5.4
GPQA Diamond (Epoch AI run)87.6%93.3%GPT-5.4
ARC-AGI72.8%93.7%GPT-5.4
ARC-AGI-217.6%74.0%GPT-5.4
SimpleBench53.2%Not measuredOnly GPT-5.1 measured
Chess puzzles42.0%48.0%GPT-5.4
Mystery game puzzles19.0%37.0%GPT-5.4
EBR-benchNot measured25.4%Only GPT-5.4 measured
VPCT58.7%Not measuredOnly GPT-5.1 measured
LMArena Text: hard prompts14571485GPT-5.4
LMArena Text: expert14701507GPT-5.4
LMArena Text: hard prompts in English14621479GPT-5.4
LMArena Vision: diagrams12641327GPT-5.4
LMArena Vision: homework12721336GPT-5.4
Kagi LLM BenchmarkNot measured63.8%Only 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 index04
Math benchmarks for GPT-5.1 and GPT-5.4
BenchmarkGPT-5.1GPT-5.4Leader
AIME93.3%96.7%GPT-5.4
MGSM (Vals AI)93.0%Not measuredOnly GPT-5.1 measured
AA Math Index94.0Not measuredOnly GPT-5.1 measured
AIME 2025 (Artificial Analysis run)94.0%Not measuredOnly GPT-5.1 measured
OTIS Mock AIME88.6%97.8%GPT-5.4
FrontierMath, tiers 1 to 3Not measured78.6%Only GPT-5.4 measured
FrontierMath, tier 4Not measured49.0%Only GPT-5.4 measured
ProofBenchNot measured56.0%Only GPT-5.4 measured
LMArena Text: math14461488GPT-5.4
LMArena Text: mathematical work14491497GPT-5.4
Neither model has a result on the other 3 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index58
Knowledge benchmarks for GPT-5.1 and GPT-5.4
BenchmarkGPT-5.1GPT-5.4Leader
MMLU Pro86.4%87.5%GPT-5.4
LegalBench (Vals AI)85.7%86.0%GPT-5.4
CaseLaw (Vals AI)73.4%63.8%GPT-5.1
CorpFin (Vals AI)63.8%65.3%GPT-5.4
TaxEval (Vals AI)74.9%74.0%GPT-5.1
MortgageTax (Vals AI)61.4%68.3%GPT-5.4
MedQA (Vals AI)96.4%96.1%GPT-5.1
MedCode (Vals AI)52.7%41.3%GPT-5.1
MMLU Pro (Artificial Analysis run)87.0%Not measuredOnly GPT-5.1 measured
SimpleQA Verified48.0%45.1%GPT-5.1
LMArena Text: science14591479GPT-5.4
LMArena Text: medicine and healthcare14561458GPT-5.4
LMArena Text: legal and government14591486GPT-5.4
LMArena Text: business and finance14311474GPT-5.4
FACTS Benchmark Suite45.0%Not measuredOnly GPT-5.1 measured
Neither model has a result on the other 3 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index26
Web design benchmarks for GPT-5.1 and GPT-5.4
BenchmarkGPT-5.1GPT-5.4Leader
Design Arena: websites11871198GPT-5.4
Design Arena: UI components11631212GPT-5.4
Design Arena: game dev11781194GPT-5.4
Design Arena: data viz12041196GPT-5.1
Design Arena: 3D10601065GPT-5.4
Design Arena: all code categories11701179GPT-5.4
Design Arena: SVG11491141GPT-5.1
Design Arena: ASCII art11191188GPT-5.4
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index132
General benchmarks for GPT-5.1 and GPT-5.4
BenchmarkGPT-5.1GPT-5.4Leader
SAGE (Vals AI)43.2%43.3%GPT-5.4
MedScribe (Vals AI)88.1%77.5%GPT-5.1
IFBench72.9%73.9%GPT-5.4
Long context reasoning (AA-LCR)80.0%82.0%GPT-5.4
DTBench90.1%94.4%GPT-5.4
LMCA43.9%52.0%GPT-5.4
CL-bench23.7%27.9%GPT-5.4
Furniture assemblyNot measured37.5%Only GPT-5.4 measured
CL-bench Life17.3%21.7%GPT-5.4
LMArena Text14431469GPT-5.4
LMArena Text: instruction following14431469GPT-5.4
LMArena Text: creative writing14271439GPT-5.4
LMArena Vision12501303GPT-5.4
LMArena Text: English14511467GPT-5.4
LMArena Text: non-English14311465GPT-5.4
LMArena Text: Chinese14951520GPT-5.4
LMArena Text: German14381472GPT-5.4
LMArena Text: Japanese14531485GPT-5.4
LMArena Text: Korean14011449GPT-5.4
LMArena Text: French14501493GPT-5.4
LMArena Text: Spanish14331452GPT-5.4
LMArena Text: Russian14351480GPT-5.4
LMArena Text: Polish14401472GPT-5.4
LMArena Text: multi-turn14501482GPT-5.4
LMArena Text: longer queries14481473GPT-5.4
LMArena Text: writing and language14281458GPT-5.4
LMArena Text: entertainment and media14151438GPT-5.4
LMArena Vision: English12521301GPT-5.4
LMArena Vision: Chinese12701354GPT-5.4
LMArena Vision: reading text12551312GPT-5.4
LMArena Vision: humor12631284GPT-5.4
LMArena Vision: captioning1252Not measuredOnly GPT-5.1 measured
LMArena Vision: recognizing things12321251GPT-5.4
LMArena Vision: creative writing12661283GPT-5.4
LMArena Document14031471GPT-5.4
Neither model has a result on the other 11 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GPT-5.1 and GPT-5.4
BenchmarkGPT-5.1GPT-5.4Leader
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-5.1$0.0063Fits in one request
GPT-5.4$0.010Fits in one request

GPT-5.1 costs less

Repository review

50K input + 3K output tokens

GPT-5.1$0.092Fits in one request
GPT-5.4$0.170Fits in one request

GPT-5.1 costs less

Agent loop

200K input + 10K output tokens

GPT-5.1$0.350Fits in one request
GPT-5.4$0.650Fits in one request

GPT-5.1 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-5.1 and GPT-5.4
SpecificationOpenAI GPT-5.1OpenAI GPT-5.4
MakerOpenAIOpenAI
WeightsNo public weights listedNo public weights listed
ReleasedNovember 13, 2025March 5, 2026
Context window400K1.05M
Input $/Mtok$1.25$2.50
Output $/Mtok$10.00$15.00
Coding index40.464.4
Independent measurements behind it33
Cost per solved task$1.02$1.02
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-5.1 better than GPT-5.4 for coding?

Pick GPT-5.4 for coding: it scores 64.4 on the Vibe Coding Index against 40.4 for GPT-5.1. Pick GPT-5.1 for output tokens 1.5x cheaper ($10.00 vs $15.00 per million) and work measured by LiveCodeBench (Vals AI) or DeepResearch Bench, where it leads.

Which is cheaper, GPT-5.1 or GPT-5.4?

GPT-5.1 costs $1.25 per million input tokens and $10.00 per million output tokens; GPT-5.4 costs $2.50 and $15.00. Per solved SWE-bench Verified task, GPT-5.1 costs $1.02 and GPT-5.4 $1.02.

Which has the bigger context window, GPT-5.1 or GPT-5.4?

GPT-5.4, with 1.05M tokens against 400K.

How many benchmarks is the GPT-5.1 vs GPT-5.4 comparison based on?

91 benchmarks have a result for both models as of October 22, 2026. GPT-5.1's coding index rests on 3 independent benchmarks and GPT-5.4's on 3. 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-5-1-vs-gpt-5-4" title="GPT-5.1 vs GPT-5.4, 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-5-1-vs-gpt-5-4">LLMBenchmark.ai, GPT-5.1 vs GPT-5.4</a></p>