Enter opens the highlighted result. Esc closes.

GPT-5 vs GPT-5.3-Codex

GPT-5.3-Codex leads GPT-5 on coding, 60.0 to 38.8 on the Vibe Coding Index, and GPT-5 costs 1.4x less per output token ($10.00 vs $14.00 per million). Updated October 21, 2026; 25 benchmarks have a result for both.

Replace GPT-5 Replace GPT-5.3-Codex
GPT-5OpenAI
Coding
38.8
Overall
51.5
Output $/M
$10.00
10
GPT-5.3-CodexOpenAI
Coding
60.0
Overall
No data
Output $/M
$14.00
15
vs boards won of 25 shared

GPT-5 or GPT-5.3-Codex: which to pick

Pick GPT-5.3-Codex for coding: it scores 60.0 on the Vibe Coding Index against 38.8 for GPT-5.

Pick GPT-5 for output tokens 1.4x cheaper ($10.00 vs $14.00 per million) and a lower cost per solved task ($0.39 against $0.59).

Model AGPT-5OpenAI38.8/1003 independent benchmarks$0.39 per solved taskModel BGPT-5.3-CodexOpenAI60.0/1003 independent benchmarks$0.59 per solved task
Shared results
25
GPT-5 only
88
GPT-5.3-Codex only
9
Like-for-like categories
6 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, GPT-5.3-Codex 60.0 from 3 benchmarks.
Widest category
Coding, 16.1 points apart on average over 9 shared benchmarks.
Widest single gap
Vibe Code Bench (Vals AI): GPT-5 20.1% against GPT-5.3-Codex 61.8%, 41.7 points apart on a 0 to 100 scale.
Output price
GPT-5 $10.00 against GPT-5.3-Codex $14.00 per million output tokens, 1.4x apart.
Context window
Both take 400K tokens.

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-5BarsGPT-5.3-CodexTies
Coding27
Intelligence02
Agentic work11
Reasoning02
Web design71
General02
All shared boards1015
CodingLike-for-likeSix of these boards feed the Vibe Coding Index27
Coding benchmarks for GPT-5 and GPT-5.3-Codex
BenchmarkGPT-5GPT-5.3-CodexLeader
Aider polyglot88.0%Not measuredOnly GPT-5 measured
SWE-bench Verified74.4%Not measuredOnly GPT-5 measured
LMArena WebDev14181407GPT-5
AA Coding Index37.8Not measuredOnly GPT-5 measured
LiveCodeBench (Vals AI)85.9%87.3%GPT-5.3-Codex
SWE-bench Verified (Vals AI)69.0%78.0%GPT-5.3-Codex
Vibe Code Bench (Vals AI)20.1%61.8%GPT-5.3-Codex
IOI (Vals AI)Not measured53.8%Only GPT-5.3-Codex measured
LiveCodeBench (Artificial Analysis run)84.6%Not measuredOnly GPT-5 measured
Terminal-Bench Hard (Artificial Analysis run)37.9%53.0%GPT-5.3-Codex
Terminal-Bench49.6%78.4%GPT-5.3-Codex
GSO-Bench6.9%Not measuredOnly GPT-5 measured
WeirdML60.7%79.3%GPT-5.3-Codex
ALE-Bench807.61655.2GPT-5.3-Codex
LMArena Text: coding1436Not measuredOnly GPT-5 measured
LMArena Text: software and IT1420Not measuredOnly GPT-5 measured
LMArena WebDev: ReactNot measured1395Only GPT-5.3-Codex measured
LMArena WebDev: HTML14271422GPT-5
LMArena WebDev: from an imageNot measured1443Only GPT-5.3-Codex measured
SciCode0.0%Not measuredOnly GPT-5 measured
Neither model has a result on the other 19 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for GPT-5 and GPT-5.3-Codex
BenchmarkGPT-5GPT-5.3-CodexLeader
AA Intelligence Index23.032.5GPT-5.3-Codex
Epoch Capabilities Index150.0157.1GPT-5.3-Codex
Agentic workLike-for-likePublic boards, shown as published, kept out of the index11
Agentic work benchmarks for GPT-5 and GPT-5.3-Codex
BenchmarkGPT-5GPT-5.3-CodexLeader
Design Arena: agents building web appsNot measured1032Only GPT-5.3-Codex measured
Design Arena: agents building full-stack appsNot measured983Only GPT-5.3-Codex measured
Design Arena: agents building mobile appsNot measured1054Only GPT-5.3-Codex measured
Design Arena: agents building Android appsNot measured1084Only GPT-5.3-Codex measured
Design Arena: agents building Godot gamesNot measured1124Only GPT-5.3-Codex measured
τ²-Bench Telecom86.5%86.0%GPT-5
METR time horizons69.6%74.5%GPT-5.3-Codex
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
Vending-Bench 2Not measured$5,940Only GPT-5.3-Codex measured
BrowseComp-Plus80.0%Not measuredOnly GPT-5 measured
Neither model has a result on the other 28 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index02
Reasoning benchmarks for GPT-5 and GPT-5.3-Codex
BenchmarkGPT-5GPT-5.3-CodexLeader
GPQA Diamond85.6%Not measuredOnly GPT-5 measured
MMMU Pro (Vals AI)81.5%Not measuredOnly GPT-5 measured
GPQA Diamond (Artificial Analysis run)85.4%91.5%GPT-5.3-Codex
Humanity's Last Exam (Artificial Analysis run)28.5%42.5%GPT-5.3-Codex
GPQA Diamond (Epoch AI run)86.2%Not measuredOnly GPT-5 measured
ARC-AGI65.7%Not measuredOnly GPT-5 measured
ARC-AGI-29.9%Not measuredOnly GPT-5 measured
SimpleBench56.7%Not measuredOnly GPT-5 measured
Chess puzzles36.0%Not measuredOnly GPT-5 measured
Mystery game puzzles23.0%Not measuredOnly GPT-5 measured
EBR-bench12.7%Not measuredOnly GPT-5 measured
VPCT66.0%Not measuredOnly GPT-5 measured
LMArena Text: hard prompts1415Not measuredOnly GPT-5 measured
LMArena Text: expert1419Not measuredOnly GPT-5 measured
LMArena Text: hard prompts in English1419Not measuredOnly GPT-5 measured
LMArena Vision: diagrams1252Not measuredOnly GPT-5 measured
LMArena Vision: homework1278Not measuredOnly GPT-5 measured
Kagi LLM Benchmark72.7%Not measuredOnly GPT-5 measured
Neither model has a result on the other 4 boards here
MathDirectional onlyPublic boards, shown as published, kept out of the index12 boards, none shared
Math benchmarks for GPT-5 and GPT-5.3-Codex
BenchmarkGPT-5GPT-5.3-CodexLeader
AIME93.4%Not measuredOnly GPT-5 measured
MATH 500 (Vals AI)96.0%Not measuredOnly GPT-5 measured
MGSM (Vals AI)92.8%Not measuredOnly GPT-5 measured
AA Math Index94.3Not measuredOnly GPT-5 measured
AIME 2025 (Artificial Analysis run)94.3%Not measuredOnly GPT-5 measured
OTIS Mock AIME91.4%Not measuredOnly GPT-5 measured
FrontierMath, tiers 1 to 355.4%Not measuredOnly GPT-5 measured
FrontierMath, tier 422.0%Not measuredOnly GPT-5 measured
ProofBench18.0%Not measuredOnly GPT-5 measured
MATH level 598.1%Not measuredOnly GPT-5 measured
LMArena Text: math1407Not measuredOnly GPT-5 measured
LMArena Text: mathematical work1405Not measuredOnly GPT-5 measured
Neither model has a result on the other 1 board here
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index16 boards, none shared
Knowledge benchmarks for GPT-5 and GPT-5.3-Codex
BenchmarkGPT-5GPT-5.3-CodexLeader
MMLU Pro86.5%Not measuredOnly GPT-5 measured
LegalBench (Vals AI)86.0%Not measuredOnly GPT-5 measured
CaseLaw (Vals AI)66.5%Not measuredOnly GPT-5 measured
CorpFin (Vals AI)61.1%Not measuredOnly GPT-5 measured
TaxEval (Vals AI)73.4%Not measuredOnly GPT-5 measured
MortgageTax (Vals AI)65.5%Not measuredOnly GPT-5 measured
MedQA (Vals AI)96.3%Not measuredOnly GPT-5 measured
MedCode (Vals AI)49.6%Not measuredOnly GPT-5 measured
MMLU Pro (Artificial Analysis run)87.1%Not measuredOnly GPT-5 measured
SimpleQA Verified50.1%Not measuredOnly GPT-5 measured
GeoBench81.0%Not measuredOnly GPT-5 measured
LMArena Text: science1416Not measuredOnly GPT-5 measured
LMArena Text: medicine and healthcare1413Not measuredOnly GPT-5 measured
LMArena Text: legal and government1431Not measuredOnly GPT-5 measured
LMArena Text: business and finance1400Not measuredOnly GPT-5 measured
FACTS Benchmark Suite56.5%Not measuredOnly GPT-5 measured
Neither model has a result on the other 2 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index71
Web design benchmarks for GPT-5 and GPT-5.3-Codex
BenchmarkGPT-5GPT-5.3-CodexLeader
Design Arena: websites11841163GPT-5
Design Arena: UI components11801137GPT-5
Design Arena: game dev11851161GPT-5
Design Arena: data viz12211164GPT-5
Design Arena: 3D10561008GPT-5
Design Arena: all code categories11681145GPT-5
Design Arena: SVG11891129GPT-5
Design Arena: ASCII art11411152GPT-5.3-Codex
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index02
General benchmarks for GPT-5 and GPT-5.3-Codex
BenchmarkGPT-5GPT-5.3-CodexLeader
SAGE (Vals AI)43.7%Not measuredOnly GPT-5 measured
MedScribe (Vals AI)83.7%Not measuredOnly GPT-5 measured
IFBench73.1%75.4%GPT-5.3-Codex
Long context reasoning (AA-LCR)78.2%83.3%GPT-5.3-Codex
DTBench90.7%Not measuredOnly GPT-5 measured
LMCA40.0%Not measuredOnly GPT-5 measured
Fiction.LiveBench97.2%Not measuredOnly GPT-5 measured
Creative writing (Lech Mazur)86.0%Not measuredOnly GPT-5 measured
LMArena Text1406Not measuredOnly GPT-5 measured
LMArena Text: instruction following1388Not measuredOnly GPT-5 measured
LMArena Text: creative writing1365Not measuredOnly GPT-5 measured
LMArena Vision1232Not measuredOnly GPT-5 measured
LMArena Text: English1409Not measuredOnly GPT-5 measured
LMArena Text: non-English1397Not measuredOnly GPT-5 measured
LMArena Text: Chinese1422Not measuredOnly GPT-5 measured
LMArena Text: German1416Not measuredOnly GPT-5 measured
LMArena Text: Japanese1409Not measuredOnly GPT-5 measured
LMArena Text: Korean1360Not measuredOnly GPT-5 measured
LMArena Text: French1409Not measuredOnly GPT-5 measured
LMArena Text: Spanish1399Not measuredOnly GPT-5 measured
LMArena Text: Russian1407Not measuredOnly GPT-5 measured
LMArena Text: Polish1431Not measuredOnly GPT-5 measured
LMArena Text: multi-turn1426Not measuredOnly GPT-5 measured
LMArena Text: longer queries1399Not measuredOnly GPT-5 measured
LMArena Text: writing and language1384Not measuredOnly GPT-5 measured
LMArena Text: entertainment and media1379Not measuredOnly GPT-5 measured
LMArena Vision: English1231Not measuredOnly GPT-5 measured
LMArena Vision: Chinese1267Not measuredOnly GPT-5 measured
LMArena Vision: reading text1246Not measuredOnly GPT-5 measured
LMArena Vision: humor1219Not measuredOnly GPT-5 measured
LMArena Vision: captioning1212Not measuredOnly GPT-5 measured
LMArena Vision: recognizing things1263Not measuredOnly GPT-5 measured
LMArena Vision: creative writing1210Not measuredOnly GPT-5 measured
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 GPT-5.3-Codex
BenchmarkGPT-5GPT-5.3-CodexLeader
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
GPT-5.3-Codex$0.0088Fits in one request

GPT-5 costs less

Repository review

50K input + 3K output tokens

GPT-5$0.092Fits in one request
GPT-5.3-Codex$0.130Fits in one request

GPT-5 costs less

Agent loop

200K input + 10K output tokens

GPT-5$0.350Fits in one request
GPT-5.3-Codex$0.490Fits in one request

GPT-5 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 GPT-5.3-Codex
SpecificationOpenAI GPT-5OpenAI GPT-5.3-Codex
MakerOpenAIOpenAI
WeightsNo public weights listedNo public weights listed
ReleasedAugust 7, 2025February 24, 2026
Context window400K400K
Input $/Mtok$1.25$1.75
Output $/Mtok$10.00$14.00
Coding index38.860.0
Independent measurements behind it33
Cost per solved task$0.39$0.59
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-5 better than GPT-5.3-Codex for coding?

Pick GPT-5.3-Codex for coding: it scores 60.0 on the Vibe Coding Index against 38.8 for GPT-5. Pick GPT-5 for output tokens 1.4x cheaper ($10.00 vs $14.00 per million) and a lower cost per solved task ($0.39 against $0.59).

Which is cheaper, GPT-5 or GPT-5.3-Codex?

GPT-5 costs $1.25 per million input tokens and $10.00 per million output tokens; GPT-5.3-Codex costs $1.75 and $14.00. Per solved SWE-bench Verified task, GPT-5 costs $0.39 and GPT-5.3-Codex $0.59.

Which has the bigger context window, GPT-5 or GPT-5.3-Codex?

Neither: both take 400K tokens.

How many benchmarks is the GPT-5 vs GPT-5.3-Codex comparison based on?

25 benchmarks have a result for both models as of October 21, 2026. GPT-5's coding index rests on 3 independent benchmarks and GPT-5.3-Codex'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-vs-gpt-5-3-codex" title="GPT-5 vs GPT-5.3-Codex, 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-vs-gpt-5-3-codex">LLMBenchmark.ai, GPT-5 vs GPT-5.3-Codex</a></p>