Enter opens the highlighted result. Esc closes.

GPT-5 vs o4 Mini

GPT-5 leads o4 Mini on the Artificial Analysis Intelligence Index, 23.0 to 16.7, and o4 Mini costs 2.3x less per output token ($4.40 vs $10.00 per million). Updated October 22, 2026; 96 benchmarks have a result for both.

Replace GPT-5 Replace o4 Mini
GPT-5OpenAI
Coding
38.8
Overall
51.5
Output $/M
$10.00
92
o4 MiniOpenAI
Coding
No data
Overall
43.0
Output $/M
$4.40
4
vs boards won of 96 shared

GPT-5 or o4 Mini: which to pick

Only GPT-5 has a coding score, 38.8 on the Vibe Coding Index; o4 Mini has none, so the two cannot be ranked on coding.

Pick o4 Mini for output tokens 2.3x cheaper ($4.40 vs $10.00 per million) and work measured by LiveCodeBench (Artificial Analysis run) or SciCode, where it leads.

Model AGPT-5OpenAI38.8/1003 independent benchmarks$0.39 per solved taskModel Bo4 MiniOpenAINot benchmarkedNo coding benchmarkNo cost per solved task
Shared results
96
GPT-5 only
17
o4 Mini only
2
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 38.8 from 3 benchmarks, o4 Mini not benchmarked.
Widest category
Web design, 24.9 points apart on average over 6 shared benchmarks.
Widest single gap
τ²-Bench Telecom: GPT-5 86.5% against o4 Mini 55.6%, 31.0 points apart on a 0 to 100 scale.
Output price
GPT-5 $10.00 against o4 Mini $4.40 per million output tokens, 2.3x apart.
Context window
GPT-5 takes 400K tokens, o4 Mini 200K.

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-5Barso4 MiniTies
Coding82
Intelligence20
Agentic work20
Reasoning170
Math101
Knowledge141
Web design60
General330
All shared boards924
CodingLike-for-likeSix of these boards feed the Vibe Coding Index82
Coding benchmarks for GPT-5 and o4 Mini
BenchmarkGPT-5o4 MiniLeader
Aider polyglot88.0%72.0%GPT-5
SWE-bench Verified74.4%64.6%GPT-5
SWE-bench MultimodalNot measured33.9%Only o4 Mini measured
LMArena WebDev1418Not measuredOnly GPT-5 measured
AA Coding Index37.8Not measuredOnly GPT-5 measured
LiveCodeBench (Vals AI)85.9%82.2%GPT-5
SWE-bench Verified (Vals AI)69.0%Not measuredOnly GPT-5 measured
Vibe Code Bench (Vals AI)20.1%Not measuredOnly GPT-5 measured
LiveCodeBench (Artificial Analysis run)84.6%85.9%o4 Mini
Terminal-Bench Hard (Artificial Analysis run)37.9%15.2%GPT-5
Terminal-Bench49.6%Not measuredOnly GPT-5 measured
GSO-Bench6.9%3.6%GPT-5
WeirdML60.7%52.6%GPT-5
ALE-Bench807.6Not measuredOnly GPT-5 measured
LMArena Text: coding14361368GPT-5
LMArena Text: software and IT14201367GPT-5
LMArena WebDev: HTML1427Not measuredOnly GPT-5 measured
SciCode0.0%10.8%o4 Mini
Neither model has a result on the other 21 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for GPT-5 and o4 Mini
BenchmarkGPT-5o4 MiniLeader
AA Intelligence Index23.016.7GPT-5
Epoch Capabilities Index150.0145.5GPT-5
Agentic workLike-for-likePublic boards, shown as published, kept out of the index20
Agentic work benchmarks for GPT-5 and o4 Mini
BenchmarkGPT-5o4 MiniLeader
τ²-Bench Telecom86.5%55.6%GPT-5
METR time horizons69.6%63.9%GPT-5
Remote Labor Index1.7%Not measuredOnly GPT-5 measured
GDPval34.8%Not measuredOnly GPT-5 measured
BALROG32.8%Not measuredOnly GPT-5 measured
DeepResearch Bench49.6%Not measuredOnly GPT-5 measured
Berkeley Function Calling Leaderboard (BFCL)Not measured53.2%Only o4 Mini measured
BrowseComp-Plus80.0%Not measuredOnly GPT-5 measured
Neither model has a result on the other 33 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index170
Reasoning benchmarks for GPT-5 and o4 Mini
BenchmarkGPT-5o4 MiniLeader
GPQA Diamond85.6%74.5%GPT-5
MMMU Pro (Vals AI)81.5%79.7%GPT-5
GPQA Diamond (Artificial Analysis run)85.4%78.4%GPT-5
Humanity's Last Exam (Artificial Analysis run)28.5%16.5%GPT-5
GPQA Diamond (Epoch AI run)86.2%79.6%GPT-5
ARC-AGI65.7%58.7%GPT-5
ARC-AGI-29.9%6.1%GPT-5
SimpleBench56.7%38.7%GPT-5
Chess puzzles36.0%22.0%GPT-5
Mystery game puzzles23.0%5.0%GPT-5
EBR-bench12.7%Not measuredOnly GPT-5 measured
VPCT66.0%57.5%GPT-5
LMArena Text: hard prompts14151351GPT-5
LMArena Text: expert14191343GPT-5
LMArena Text: hard prompts in English14191366GPT-5
LMArena Vision: diagrams12521197GPT-5
LMArena Vision: homework12781234GPT-5
Kagi LLM Benchmark72.7%67.6%GPT-5
Neither model has a result on the other 4 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index101
Math benchmarks for GPT-5 and o4 Mini
BenchmarkGPT-5o4 MiniLeader
AIME93.4%83.7%GPT-5
MATH 500 (Vals AI)96.0%94.2%GPT-5
MGSM (Vals AI)92.8%93.4%o4 Mini
AA Math Index94.390.7GPT-5
AIME 2025 (Artificial Analysis run)94.3%90.7%GPT-5
OTIS Mock AIME91.4%81.7%GPT-5
FrontierMath, tiers 1 to 355.4%36.1%GPT-5
FrontierMath, tier 422.0%4.9%GPT-5
ProofBench18.0%Not measuredOnly GPT-5 measured
MATH level 598.1%97.8%GPT-5
LMArena Text: math14071389GPT-5
LMArena Text: mathematical work14051379GPT-5
Neither model has a result on the other 1 board here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index141
Knowledge benchmarks for GPT-5 and o4 Mini
BenchmarkGPT-5o4 MiniLeader
MMLU Pro86.5%80.6%GPT-5
LegalBench (Vals AI)86.0%79.2%GPT-5
CaseLaw (Vals AI)66.5%Not measuredOnly GPT-5 measured
CorpFin (Vals AI)61.1%59.0%GPT-5
TaxEval (Vals AI)73.4%74.8%o4 Mini
MortgageTax (Vals AI)65.5%64.8%GPT-5
MedQA (Vals AI)96.3%96.0%GPT-5
MedCode (Vals AI)49.6%33.8%GPT-5
MMLU Pro (Artificial Analysis run)87.1%83.2%GPT-5
SimpleQA Verified50.1%19.6%GPT-5
GeoBench81.0%64.0%GPT-5
LMArena Text: science14161354GPT-5
LMArena Text: medicine and healthcare14131361GPT-5
LMArena Text: legal and government14311350GPT-5
LMArena Text: business and finance14001325GPT-5
FACTS Benchmark Suite56.5%28.1%GPT-5
Neither model has a result on the other 2 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index60
Web design benchmarks for GPT-5 and o4 Mini
BenchmarkGPT-5o4 MiniLeader
Design Arena: websites1184985GPT-5
Design Arena: UI components1177983GPT-5
Design Arena: game dev11851003GPT-5
Design Arena: data viz1221991GPT-5
Design Arena: 3D1055854GPT-5
Design Arena: all code categories1167976GPT-5
Design Arena: SVG1188Not measuredOnly GPT-5 measured
Design Arena: ASCII art1141Not measuredOnly GPT-5 measured
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index330
General benchmarks for GPT-5 and o4 Mini
BenchmarkGPT-5o4 MiniLeader
SAGE (Vals AI)43.7%41.1%GPT-5
MedScribe (Vals AI)83.7%69.1%GPT-5
IFBench73.1%68.7%GPT-5
Long context reasoning (AA-LCR)78.2%61.0%GPT-5
DTBench90.7%77.6%GPT-5
LMCA40.0%26.5%GPT-5
Fiction.LiveBench97.2%77.8%GPT-5
Creative writing (Lech Mazur)86.0%75.0%GPT-5
LMArena Text14061353GPT-5
LMArena Text: instruction following13881320GPT-5
LMArena Text: creative writing13651294GPT-5
LMArena Vision12321194GPT-5
LMArena Text: English14091367GPT-5
LMArena Text: non-English13971337GPT-5
LMArena Text: Chinese14221354GPT-5
LMArena Text: German14161336GPT-5
LMArena Text: Japanese14091308GPT-5
LMArena Text: Korean13601312GPT-5
LMArena Text: French14091364GPT-5
LMArena Text: Spanish13991346GPT-5
LMArena Text: Russian14071334GPT-5
LMArena Text: Polish14311358GPT-5
LMArena Text: multi-turn14261350GPT-5
LMArena Text: longer queries13991315GPT-5
LMArena Text: writing and language13841316GPT-5
LMArena Text: entertainment and media13791308GPT-5
LMArena Vision: English12311200GPT-5
LMArena Vision: Chinese12671218GPT-5
LMArena Vision: reading text12461197GPT-5
LMArena Vision: humor12191199GPT-5
LMArena Vision: captioning12121204GPT-5
LMArena Vision: recognizing things12631218GPT-5
LMArena Vision: creative writing12101166GPT-5
Neither model has a result on the other 13 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GPT-5 and o4 Mini
BenchmarkGPT-5o4 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$0.0063Fits in one request
o4 Mini$0.0033Fits in one request

o4 Mini costs less

Repository review

50K input + 3K output tokens

GPT-5$0.092Fits in one request
o4 Mini$0.068Fits in one request

o4 Mini costs less

Agent loop

200K input + 10K output tokens

GPT-5$0.350Fits in one request
o4 Mini$0.264Exceeds the context window

o4 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 and o4 Mini
SpecificationOpenAI GPT-5OpenAI o4 Mini
MakerOpenAIOpenAI
WeightsNo public weights listedNo public weights listed
ReleasedAugust 7, 2025April 16, 2025
Context window400K200K
Input $/Mtok$1.25$1.10
Output $/Mtok$10.00$4.40
Coding index38.8Not benchmarked
Independent measurements behind it30
Cost per solved task$0.39Not computable
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-5 better than o4 Mini for coding?

Only GPT-5 has a coding score, 38.8 on the Vibe Coding Index; o4 Mini has none, so the two cannot be ranked on coding. Pick o4 Mini for output tokens 2.3x cheaper ($4.40 vs $10.00 per million) and work measured by LiveCodeBench (Artificial Analysis run) or SciCode, where it leads.

Which is cheaper, GPT-5 or o4 Mini?

GPT-5 costs $1.25 per million input tokens and $10.00 per million output tokens; o4 Mini costs $1.10 and $4.40. Cost per solved task is known for GPT-5 only.

Which has the bigger context window, GPT-5 or o4 Mini?

GPT-5, with 400K tokens against 200K.

How many benchmarks is the GPT-5 vs o4 Mini comparison based on?

96 benchmarks have a result for both models as of October 22, 2026. GPT-5's coding index rests on 3 independent benchmarks and o4 Mini's on 0. Where only one model has a result, the page says so instead of naming a winner.

Compare next