Enter opens the highlighted result. Esc closes.

GPT-5.4 Mini vs GPT-5.5

GPT-5.5 leads GPT-5.4 Mini on coding, 71.3 to 35.8 on the Vibe Coding Index, and GPT-5.4 Mini costs 6.7x less per output token ($4.50 vs $30.00 per million). Updated October 21, 2026; 87 benchmarks have a result for both.

Replace GPT-5.4 Mini Replace GPT-5.5
GPT-5.4 MiniOpenAI
Coding
35.8
Overall
44.0
Output $/M
$4.50
1
GPT-5.5OpenAI
Coding
71.3
Overall
69.2
Output $/M
$30.00
86
vs boards won of 87 shared

GPT-5.4 Mini or GPT-5.5: which to pick

Pick GPT-5.5 for coding: it scores 71.3 on the Vibe Coding Index against 35.8 for GPT-5.4 Mini.

Pick GPT-5.4 Mini for output tokens 6.7x cheaper ($4.50 vs $30.00 per million) and a lower cost per solved task ($0.70 against $1.65).

Model AGPT-5.4 MiniOpenAI35.8/1005 independent benchmarks$0.70 per solved taskModel BGPT-5.5OpenAI71.3/1005 independent benchmarks$1.65 per solved task
Shared results
87
GPT-5.4 Mini only
2
GPT-5.5 only
47
Like-for-like categories
7 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 Mini 35.8 from 5 benchmarks, GPT-5.5 71.3 from 5 benchmarks.
Widest category
Math, 25.5 points apart on average over 6 shared benchmarks.
Widest single gap
ARC-AGI-2: GPT-5.4 Mini 18.9% against GPT-5.5 85.0%, 66.1 points apart on a 0 to 100 scale.
Output price
GPT-5.4 Mini $4.50 against GPT-5.5 $30.00 per million output tokens, 6.7x apart.
Context window
GPT-5.4 Mini takes 400K tokens, GPT-5.5 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-5.4 MiniBarsGPT-5.5Ties
Coding017
Intelligence02
Agentic work07
Reasoning015
Math15
Knowledge010
General030
All shared boards186
CodingLike-for-likeSix of these boards feed the Vibe Coding Index017
Coding benchmarks for GPT-5.4 Mini and GPT-5.5
BenchmarkGPT-5.4 MiniGPT-5.5Leader
LiveBench codingNot measured85.1%Only GPT-5.5 measured
LMArena WebDev13971513GPT-5.5
AA Coding Index56.174.9GPT-5.5
LiveCodeBench (Vals AI)81.5%85.3%GPT-5.5
SWE-bench Verified (Vals AI)73.0%82.6%GPT-5.5
Terminal-Bench 4 (Vals AI)2.5%Not measuredOnly GPT-5.4 Mini measured
Vibe Code Bench (Vals AI)48.0%69.8%GPT-5.5
Code migration (Vals AI)12.9%45.2%GPT-5.5
ProgramBench (Vals AI)0.0%0.5%GPT-5.5
Terminal-Bench 2.1 (Vals AI)54.7%76.4%GPT-5.5
SkillsBench (Vals AI)Not measured62.5%Only GPT-5.5 measured
SciCode52.1%56.1%GPT-5.5
Terminal-Bench Hard (Artificial Analysis run)52.3%60.6%GPT-5.5
Terminal-BenchNot measured84.7%Only GPT-5.5 measured
FrontierCode27.0%43.0%GPT-5.5
GSO-BenchNot measured40.2%Only GPT-5.5 measured
WeirdML60.3%84.9%GPT-5.5
MirrorCodeNot measured10.0%Only GPT-5.5 measured
ALE-Bench1188.61943.0GPT-5.5
LMArena Text: coding14381494GPT-5.5
LMArena Text: software and IT14361493GPT-5.5
LMArena WebDev: React13861506GPT-5.5
LMArena WebDev: HTML14191543GPT-5.5
LMArena WebDev: from an imageNot measured1532Only GPT-5.5 measured
Neither model has a result on the other 15 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for GPT-5.4 Mini and GPT-5.5
BenchmarkGPT-5.4 MiniGPT-5.5Leader
AA Intelligence Index24.138.4GPT-5.5
Epoch Capabilities Index148.7159.7GPT-5.5
Agentic workLike-for-likePublic boards, shown as published, kept out of the index07
Agentic work benchmarks for GPT-5.4 Mini and GPT-5.5
BenchmarkGPT-5.4 MiniGPT-5.5Leader
AA Agentic Index17.936.4GPT-5.5
Design Arena: agents building web appsNot measured1117Only GPT-5.5 measured
Design Arena: agents building full-stack appsNot measured1089Only GPT-5.5 measured
Design Arena: agents building mobile appsNot measured1143Only GPT-5.5 measured
Design Arena: agents building Android appsNot measured1163Only GPT-5.5 measured
Design Arena: agents building Godot gamesNot measured1173Only GPT-5.5 measured
Design Arena: agents building gamesNot measured1179Only GPT-5.5 measured
Design Arena: HTML slidesNot measured1067Only GPT-5.5 measured
Design Arena: PowerPoint slidesNot measured1109Only GPT-5.5 measured
LiveBench researchNot measured61.6%Only GPT-5.5 measured
LiveBench browserNot measured48.0%Only GPT-5.5 measured
LiveBench long agenticNot measured44.6%Only GPT-5.5 measured
SREBench (Vals AI)Not measured3.8%Only GPT-5.5 measured
Finance Agent (Vals AI)45.4%51.8%GPT-5.5
Tax Agent Bench (Vals AI)8.8%21.8%GPT-5.5
Legal Research Bench (Vals AI)12.5%40.4%GPT-5.5
Excel Modeling Benchmark (Vals AI)45.4%64.5%GPT-5.5
Time Horizon Index (Vals AI)Not measured8.3%Only GPT-5.5 measured
τ²-Bench Telecom83.3%93.9%GPT-5.5
APEX-AgentsNot measured55.1%Only GPT-5.5 measured
Remote Labor IndexNot measured6.3%Only GPT-5.5 measured
OSWorld 2.0Not measured13.0%Only GPT-5.5 measured
DeepResearch Bench36.3%54.0%GPT-5.5
PostTrainBenchNot measured27.2%Only GPT-5.5 measured
Vending-Bench 2Not measured$7,524Only GPT-5.5 measured
Neither model has a result on the other 16 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index015
Reasoning benchmarks for GPT-5.4 Mini and GPT-5.5
BenchmarkGPT-5.4 MiniGPT-5.5Leader
GPQA Diamond83.1%93.2%GPT-5.5
LiveBench reasoningNot measured89.7%Only GPT-5.5 measured
MMMU Pro (Vals AI)79.3%88.3%GPT-5.5
GPQA Diamond (Artificial Analysis run)87.5%93.5%GPT-5.5
Humanity's Last Exam (Artificial Analysis run)28.1%45.8%GPT-5.5
GPQA Diamond (Epoch AI run)86.9%90.7%GPT-5.5
ARC-AGI63.7%95.0%GPT-5.5
ARC-AGI-218.9%85.0%GPT-5.5
SimpleBenchNot measured69.0%Only GPT-5.5 measured
Chess puzzles23.0%63.0%GPT-5.5
Mystery game puzzles11.0%56.0%GPT-5.5
EBR-benchNot measured34.3%Only GPT-5.5 measured
LMArena Text: hard prompts14241489GPT-5.5
LMArena Text: expert14351507GPT-5.5
LMArena Text: hard prompts in English14201483GPT-5.5
LMArena Vision: diagrams12661325GPT-5.5
LMArena Vision: homework12851332GPT-5.5
Kagi LLM Benchmark37.9%88.8%GPT-5.5
Neither model has a result on the other 4 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index15
Math benchmarks for GPT-5.4 Mini and GPT-5.5
BenchmarkGPT-5.4 MiniGPT-5.5Leader
AIME95.6%Not measuredOnly GPT-5.4 Mini measured
OTIS Mock AIME88.9%84.4%GPT-5.4 Mini
FrontierMath, tiers 1 to 351.2%85.3%GPT-5.5
FrontierMath, tier 49.8%72.5%GPT-5.5
ProofBench21.0%50.0%GPT-5.5
LMArena Text: math14181487GPT-5.5
LMArena Text: mathematical work14291496GPT-5.5
Neither model has a result on the other 6 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index010
Knowledge benchmarks for GPT-5.4 Mini and GPT-5.5
BenchmarkGPT-5.4 MiniGPT-5.5Leader
MMLU Pro84.5%88.1%GPT-5.5
LegalBench (Vals AI)Not measured86.5%Only GPT-5.5 measured
CaseLaw (Vals AI)51.7%66.2%GPT-5.5
CorpFin (Vals AI)60.9%68.4%GPT-5.5
TaxEval (Vals AI)71.2%75.0%GPT-5.5
MortgageTax (Vals AI)63.5%68.8%GPT-5.5
MedCode (Vals AI)Not measured49.1%Only GPT-5.5 measured
SimpleQA Verified29.4%63.0%GPT-5.5
LMArena Text: science14171485GPT-5.5
LMArena Text: medicine and healthcare13951465GPT-5.5
LMArena Text: legal and government14141491GPT-5.5
LMArena Text: business and finance14181478GPT-5.5
FACTS Benchmark SuiteNot measured66.7%Only GPT-5.5 measured
Neither model has a result on the other 5 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index9 boards, none shared
Web design benchmarks for GPT-5.4 Mini and GPT-5.5
BenchmarkGPT-5.4 MiniGPT-5.5Leader
Design Arena: websitesNot measured1256Only GPT-5.5 measured
Design Arena: UI componentsNot measured1260Only GPT-5.5 measured
Design Arena: game devNot measured1294Only GPT-5.5 measured
Design Arena: data vizNot measured1265Only GPT-5.5 measured
Design Arena: 3DNot measured1198Only GPT-5.5 measured
Design Arena: all code categoriesNot measured1260Only GPT-5.5 measured
Design Arena: SVGNot measured1234Only GPT-5.5 measured
Design Arena: ASCII artNot measured1255Only GPT-5.5 measured
LiveBench designNot measured70.3%Only GPT-5.5 measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index030
General benchmarks for GPT-5.4 Mini and GPT-5.5
BenchmarkGPT-5.4 MiniGPT-5.5Leader
LiveBench data analysisNot measured81.6%Only GPT-5.5 measured
Public Benefits Bench (Vals AI)Not measured60.9%Only GPT-5.5 measured
SAGE (Vals AI)50.8%51.5%GPT-5.5
MedScribe (Vals AI)Not measured86.9%Only GPT-5.5 measured
Vals Multimodal Index54.2%68.1%GPT-5.5
IFBench73.3%75.9%GPT-5.5
Long context reasoning (AA-LCR)77.0%84.3%GPT-5.5
DTBench80.0%96.0%GPT-5.5
LMCA40.8%54.3%GPT-5.5
Surface Evolver BenchNot measured88.1%Only GPT-5.5 measured
Furniture assemblyNot measured44.2%Only GPT-5.5 measured
ExploitBenchNot measured47.4%Only GPT-5.5 measured
CL-bench LifeNot measured22.2%Only GPT-5.5 measured
LMArena Text14121471GPT-5.5
LMArena Text: instruction following14051478GPT-5.5
LMArena Text: creative writing13701455GPT-5.5
LMArena Vision12451297GPT-5.5
LMArena Text: English14121470GPT-5.5
LMArena Text: non-English14051467GPT-5.5
LMArena Text: Chinese14461534GPT-5.5
LMArena Text: German14081480GPT-5.5
LMArena Text: Japanese13751499GPT-5.5
LMArena Text: Korean13701460GPT-5.5
LMArena Text: French14411485GPT-5.5
LMArena Text: Spanish14051468GPT-5.5
LMArena Text: Russian14171473GPT-5.5
LMArena Text: Polish14151465GPT-5.5
LMArena Text: multi-turn14291476GPT-5.5
LMArena Text: longer queries14081484GPT-5.5
LMArena Text: writing and language13931466GPT-5.5
LMArena Text: entertainment and media13731449GPT-5.5
LMArena Vision: English12441292GPT-5.5
LMArena Vision: Chinese12861360GPT-5.5
LMArena Vision: reading text12531306GPT-5.5
LMArena Vision: humor12351293GPT-5.5
LMArena Vision: recognizing things12181231GPT-5.5
LMArena Vision: creative writing12221301GPT-5.5
LMArena DocumentNot measured1486Only GPT-5.5 measured
Neither model has a result on the other 8 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GPT-5.4 Mini and GPT-5.5
BenchmarkGPT-5.4 MiniGPT-5.5Leader
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 Mini$0.0030Fits in one request
GPT-5.5$0.020Fits in one request

GPT-5.4 Mini costs less

Repository review

50K input + 3K output tokens

GPT-5.4 Mini$0.051Fits in one request
GPT-5.5$0.340Fits in one request

GPT-5.4 Mini costs less

Agent loop

200K input + 10K output tokens

GPT-5.4 Mini$0.195Fits in one request
GPT-5.5$1.30Fits in one request

GPT-5.4 Mini 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 Mini and GPT-5.5
SpecificationOpenAI GPT-5.4 MiniOpenAI GPT-5.5
MakerOpenAIOpenAI
WeightsNo public weights listedNo public weights listed
ReleasedMarch 17, 2026April 24, 2026
Context window400K1.05M
Input $/Mtok$0.75$5.00
Output $/Mtok$4.50$30.00
Coding index35.871.3
Independent measurements behind it55
Cost per solved task$0.70$1.65
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-5.4 Mini better than GPT-5.5 for coding?

Pick GPT-5.5 for coding: it scores 71.3 on the Vibe Coding Index against 35.8 for GPT-5.4 Mini. Pick GPT-5.4 Mini for output tokens 6.7x cheaper ($4.50 vs $30.00 per million) and a lower cost per solved task ($0.70 against $1.65).

Which is cheaper, GPT-5.4 Mini or GPT-5.5?

GPT-5.4 Mini costs $0.75 per million input tokens and $4.50 per million output tokens; GPT-5.5 costs $5.00 and $30.00. Per solved SWE-bench Verified task, GPT-5.4 Mini costs $0.70 and GPT-5.5 $1.65.

Which has the bigger context window, GPT-5.4 Mini or GPT-5.5?

GPT-5.5, with 1.05M tokens against 400K.

How many benchmarks is the GPT-5.4 Mini vs GPT-5.5 comparison based on?

87 benchmarks have a result for both models as of October 21, 2026. GPT-5.4 Mini's coding index rests on 5 independent benchmarks and GPT-5.5'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-mini-vs-gpt-5-5" title="GPT-5.4 Mini vs GPT-5.5, 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-mini-vs-gpt-5-5">LLMBenchmark.ai, GPT-5.4 Mini vs GPT-5.5</a></p>