LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

GPT-5.4 vs GPT-5 Mini

Updated October 21, 2026. 84 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace GPT-5.4 Replace GPT-5 Mini
GPT-5.4OpenAI
Coding
37.7
Overall
39.0
Output $/M
$15.00
78
GPT-5 MiniOpenAI
Coding
59.8
Overall
20.6
Output $/M
$2.00
6
vs boards won of 84 shared

Decision reading

Pick GPT-5 Mini for coding: it scores 59.8 on the Vibe Coding Index against 37.7 for GPT-5.4.

Pick GPT-5.4 when the context window is larger, 1.05M against 400K tokens, or when what matters is SWE-bench Verified (Vals AI) or Vibe Code Bench (Vals AI), where it leads.

Model AGPT-5.4OpenAI37.7/1001 independent benchmark$1.02 per solved taskModel BGPT-5 MiniOpenAI59.8/1001 independent benchmark$0.08 per solved task
Shared results
84
GPT-5.4 only
25
GPT-5 Mini only
14
Like-for-like categories
8 of 9

Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.

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 37.7 from 1 benchmark, GPT-5 Mini 59.8 from 1 benchmark.
Widest category
Agentic work, 27.2 points apart on average over 2 shared benchmarks.
Widest single gap
ARC-AGI-2: GPT-5.4 74.0% against GPT-5 Mini 4.4%, 69.5 points apart on a 0 to 100 scale.
Output price
GPT-5.4 $15.00 against GPT-5 Mini $2.00 per million output tokens, 7.5x apart.
Context window
GPT-5.4 takes 1.05M tokens, GPT-5 Mini 400K.

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-5.4BarsGPT-5 MiniTies
Coding71
Intelligence20
Agentic work20
Reasoning140
Math70
Knowledge103
Web design80
General282
All shared boards786
CodingLike-for-likeFour of these boards feed the Vibe Coding Index71
Coding benchmarks for GPT-5.4 and GPT-5 Mini
BenchmarkGPT-5.4GPT-5 MiniLeader
SWE-bench VerifiedNot measured59.8%Only GPT-5 Mini measured
SWE-bench MultilingualNot measured39.7%Only GPT-5 Mini measured
LMArena WebDev1464Not measuredOnly GPT-5.4 measured
AA Coding Index71.1Not measuredOnly GPT-5.4 measured
LiveCodeBench (Vals AI)84.1%86.6%GPT-5 Mini
SWE-bench Verified (Vals AI)78.2%60.8%GPT-5.4
Vibe Code Bench (Vals AI)67.4%14.2%GPT-5.4
Code migration (Vals AI)35.0%Not measuredOnly GPT-5.4 measured
ProgramBench (Vals AI)0.5%Not measuredOnly GPT-5.4 measured
SkillsBench (Vals AI)51.7%Not measuredOnly GPT-5.4 measured
LiveCodeBench (Artificial Analysis run)Not measured83.8%Only GPT-5 Mini measured
SciCodeNot measured39.0%Only GPT-5 Mini measured
Terminal-Bench Hard (Artificial Analysis run)57.6%33.3%GPT-5.4
Terminal-Bench81.8%34.8%GPT-5.4
GSO-Bench31.4%Not measuredOnly GPT-5.4 measured
WeirdML77.7%52.7%GPT-5.4
MirrorCode15.6%Not measuredOnly GPT-5.4 measured
ALE-Bench1607.0Not measuredOnly GPT-5.4 measured
LMArena Text: coding14961406GPT-5.4
LMArena Text: software and IT14911394GPT-5.4
LMArena WebDev: React1454Not measuredOnly GPT-5.4 measured
LMArena WebDev: HTML1478Not measuredOnly GPT-5.4 measured
LMArena WebDev: from an image1436Not measuredOnly GPT-5.4 measured
Neither model has a result on the other 15 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for GPT-5.4 and GPT-5 Mini
BenchmarkGPT-5.4GPT-5 MiniLeader
AA Intelligence Index39.020.6GPT-5.4
Epoch Capabilities Index157.3145.3GPT-5.4
Agentic workLike-for-likePublic boards, shown as published, kept out of the index20
Agentic work benchmarks for GPT-5.4 and GPT-5 Mini
BenchmarkGPT-5.4GPT-5 MiniLeader
Design Arena: agents building web apps1059Not measuredOnly GPT-5.4 measured
Design Arena: agents building full-stack apps1017Not measuredOnly GPT-5.4 measured
Design Arena: agents building mobile apps1087Not measuredOnly GPT-5.4 measured
Design Arena: agents building Android apps1037Not measuredOnly GPT-5.4 measured
Design Arena: agents building Godot games1135Not measuredOnly GPT-5.4 measured
τ²-Bench Telecom87.1%71.1%GPT-5.4
APEX-Agents52.4%Not measuredOnly GPT-5.4 measured
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$6,144-$31GPT-5.4
Neither model has a result on the other 28 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index140
Reasoning benchmarks for GPT-5.4 and GPT-5 Mini
BenchmarkGPT-5.4GPT-5 MiniLeader
GPQA Diamond91.7%80.3%GPT-5.4
MMMU Pro (Vals AI)87.5%78.9%GPT-5.4
GPQA Diamond (Artificial Analysis run)92.0%82.8%GPT-5.4
Humanity's Last Exam (Artificial Analysis run)43.7%21.5%GPT-5.4
GPQA Diamond (Epoch AI run)93.3%75.0%GPT-5.4
ARC-AGI93.7%54.3%GPT-5.4
ARC-AGI-274.0%4.4%GPT-5.4
Chess puzzles48.0%24.0%GPT-5.4
Mystery game puzzles37.0%10.0%GPT-5.4
EBR-bench25.4%Not measuredOnly GPT-5.4 measured
VPCTNot measured40.2%Only GPT-5 Mini measured
LMArena Text: hard prompts14851380GPT-5.4
LMArena Text: expert15071379GPT-5.4
LMArena Text: hard prompts in English14791383GPT-5.4
LMArena Vision: diagrams13271202GPT-5.4
LMArena Vision: homework13361210GPT-5.4
Neither model has a result on the other 5 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index70
Math benchmarks for GPT-5.4 and GPT-5 Mini
BenchmarkGPT-5.4GPT-5 MiniLeader
AIME96.7%91.5%GPT-5.4
MATH 500 (Vals AI)Not measured94.8%Only GPT-5 Mini measured
MGSM (Vals AI)Not measured92.6%Only GPT-5 Mini measured
AA Math IndexNot measured90.7Only GPT-5 Mini measured
AIME 2025 (Artificial Analysis run)Not measured90.7%Only GPT-5 Mini measured
OTIS Mock AIME97.8%86.7%GPT-5.4
FrontierMath, tiers 1 to 378.6%46.7%GPT-5.4
FrontierMath, tier 449.0%12.2%GPT-5.4
ProofBench56.0%9.0%GPT-5.4
MATH level 5Not measured97.8%Only GPT-5 Mini measured
LMArena Text: math14881378GPT-5.4
LMArena Text: mathematical work14971380GPT-5.4
Neither model has a result on the other 1 board here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index103
Knowledge benchmarks for GPT-5.4 and GPT-5 Mini
BenchmarkGPT-5.4GPT-5 MiniLeader
MMLU Pro87.5%82.2%GPT-5.4
LegalBench (Vals AI)86.0%81.8%GPT-5.4
CaseLaw (Vals AI)63.8%68.5%GPT-5 Mini
CorpFin (Vals AI)65.3%60.2%GPT-5.4
TaxEval (Vals AI)74.0%75.2%GPT-5 Mini
MortgageTax (Vals AI)68.3%66.9%GPT-5.4
MedQA (Vals AI)96.1%96.1%GPT-5.4
MedCode (Vals AI)41.3%43.0%GPT-5 Mini
MMLU Pro (Artificial Analysis run)Not measured83.7%Only GPT-5 Mini measured
SimpleQA Verified45.1%21.6%GPT-5.4
LMArena Text: science14791371GPT-5.4
LMArena Text: medicine and healthcare14581380GPT-5.4
LMArena Text: legal and government14861375GPT-5.4
LMArena Text: business and finance14741358GPT-5.4
Neither model has a result on the other 3 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index80
Web design benchmarks for GPT-5.4 and GPT-5 Mini
BenchmarkGPT-5.4GPT-5 MiniLeader
Design Arena: websites11991125GPT-5.4
Design Arena: UI components12151106GPT-5.4
Design Arena: game dev11951130GPT-5.4
Design Arena: data viz11961131GPT-5.4
Design Arena: 3D10651037GPT-5.4
Design Arena: all code categories11801115GPT-5.4
Design Arena: SVG11411091GPT-5.4
Design Arena: ASCII art11881124GPT-5.4
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index282
General benchmarks for GPT-5.4 and GPT-5 Mini
BenchmarkGPT-5.4GPT-5 MiniLeader
SAGE (Vals AI)43.3%43.0%GPT-5.4
MedScribe (Vals AI)77.5%80.6%GPT-5 Mini
IFBench73.9%75.4%GPT-5 Mini
Long context reasoning (AA-LCR)82.0%72.3%GPT-5.4
DTBench94.4%80.5%GPT-5.4
LMCA52.0%34.2%GPT-5.4
Fiction.LiveBenchNot measured69.4%Only GPT-5 Mini measured
Creative writing (Lech Mazur)Not measured83.1%Only GPT-5 Mini measured
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 Text14691373GPT-5.4
LMArena Text: instruction following14691357GPT-5.4
LMArena Text: creative writing14391325GPT-5.4
LMArena Vision13031202GPT-5.4
LMArena Text: English14671379GPT-5.4
LMArena Text: non-English14651363GPT-5.4
LMArena Text: Chinese15201386GPT-5.4
LMArena Text: German14721366GPT-5.4
LMArena Text: Japanese14851341GPT-5.4
LMArena Text: Korean14491308GPT-5.4
LMArena Text: French14931386GPT-5.4
LMArena Text: Spanish14521354GPT-5.4
LMArena Text: Russian14801362GPT-5.4
LMArena Text: Polish14721378GPT-5.4
LMArena Text: multi-turn14821363GPT-5.4
LMArena Text: longer queries14731355GPT-5.4
LMArena Text: writing and language14581344GPT-5.4
LMArena Text: entertainment and media14381340GPT-5.4
LMArena Vision: English13011216GPT-5.4
LMArena Vision: Chinese13541214GPT-5.4
LMArena Vision: reading text13121199GPT-5.4
LMArena Vision: humor12841205GPT-5.4
LMArena Vision: captioningNot measured1217Only GPT-5 Mini measured
LMArena Vision: recognizing things12511226GPT-5.4
LMArena Vision: creative writing12831181GPT-5.4
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 GPT-5 Mini
BenchmarkGPT-5.4GPT-5 MiniLeader
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
GPT-5 Mini$0.0013Fits in one request

GPT-5 Mini costs less

Repository review

50K input + 3K output tokens

GPT-5.4$0.170Fits in one request
GPT-5 Mini$0.018Fits in one request

GPT-5 Mini costs less

Agent loop

200K input + 10K output tokens

GPT-5.4$0.650Fits in one request
GPT-5 Mini$0.070Fits in one request

GPT-5 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 and GPT-5 Mini
SpecificationOpenAI GPT-5.4OpenAI GPT-5 Mini
MakerOpenAIOpenAI
WeightsNo public weights listedNo public weights listed
ReleasedMarch 5, 2026August 7, 2025
Context window1.05M400K
Input $/Mtok$2.50$0.25
Output $/Mtok$15.00$2.00
Coding index37.759.8
Independent measurements behind it11
Cost per solved task$1.02$0.08
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-5.4 better than GPT-5 Mini for coding?

Pick GPT-5 Mini for coding: it scores 59.8 on the Vibe Coding Index against 37.7 for GPT-5.4. Pick GPT-5.4 when the context window is larger, 1.05M against 400K tokens, or when what matters is SWE-bench Verified (Vals AI) or Vibe Code Bench (Vals AI), where it leads.

Which is cheaper, GPT-5.4 or GPT-5 Mini?

GPT-5.4 costs $2.50 per million input tokens and $15.00 per million output tokens; GPT-5 Mini costs $0.25 and $2.00. Per solved SWE-bench Verified task, GPT-5.4 costs $1.02 and GPT-5 Mini $0.08.

Which has the bigger context window, GPT-5.4 or GPT-5 Mini?

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

How many benchmarks is the GPT-5.4 vs GPT-5 Mini comparison based on?

84 benchmarks have a result for both models as of October 21, 2026. GPT-5.4's coding index rests on 1 independent benchmark and GPT-5 Mini's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next