LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

GPT-5 vs GPT-5.2

Updated October 21, 2026. 101 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace GPT-5 Replace GPT-5.2
GPT-5OpenAI
Coding
72.2
Overall
23.0
Output $/M
$10.00
20
GPT-5.2OpenAI
Coding
58.9
Overall
30.4
Output $/M
$14.00
80
vs boards won of 101 shared, 1 tied

Decision reading

Pick GPT-5 for coding: it scores 72.2 on the Vibe Coding Index against 58.9 for GPT-5.2.

Pick GPT-5.2 when what matters is SWE-bench Verified (Vals AI) or Vibe Code Bench (Vals AI), where it leads.

Model AGPT-5OpenAI72.2/1003 independent benchmarks$0.39 per solved taskModel BGPT-5.2OpenAI58.9/1002 independent benchmarks$1.91 per solved task
Shared results
101
GPT-5 only
8
GPT-5.2 only
10
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 72.2 from 3 benchmarks, GPT-5.2 58.9 from 2 benchmarks.
Widest category
Reasoning, 9.2 points apart on average over 17 shared benchmarks.
Widest single gap
ARC-AGI-2: GPT-5 9.9% against GPT-5.2 52.9%, 43.1 points apart on a 0 to 100 scale.
Output price
GPT-5 $10.00 against GPT-5.2 $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-5.2Ties
Coding311
Intelligence02
Agentic work23
Reasoning2141 tie
Math19
Knowledge68
Web design26
General427
All shared boards20801 tie
CodingLike-for-likeFour of these boards feed the Vibe Coding Index311
Coding benchmarks for GPT-5 and GPT-5.2
BenchmarkGPT-5GPT-5.2Leader
Aider polyglot88.0%Not measuredOnly GPT-5 measured
SWE-bench Verified74.4%72.8%GPT-5
SWE-bench MultilingualNot measured66.7%Only GPT-5.2 measured
LMArena WebDev14181417GPT-5
AA Coding Index37.8Not measuredOnly GPT-5 measured
LiveCodeBench (Vals AI)85.9%85.4%GPT-5
SWE-bench Verified (Vals AI)69.0%75.8%GPT-5.2
Vibe Code Bench (Vals AI)20.1%53.5%GPT-5.2
LiveCodeBench (Artificial Analysis run)84.6%89.4%GPT-5.2
Terminal-Bench Hard (Artificial Analysis run)37.9%47.0%GPT-5.2
Terminal-Bench49.6%64.9%GPT-5.2
GSO-Bench6.9%27.5%GPT-5.2
WeirdML60.7%72.2%GPT-5.2
ALE-Bench807.61293.5GPT-5.2
LMArena Text: coding14361442GPT-5.2
LMArena Text: software and IT14201434GPT-5.2
LMArena WebDev: HTML14271427GPT-5.2
Neither model has a result on the other 21 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for GPT-5 and GPT-5.2
BenchmarkGPT-5GPT-5.2Leader
AA Intelligence Index23.030.4GPT-5.2
Epoch Capabilities Index150.0153.7GPT-5.2
Agentic workLike-for-likePublic boards, shown as published, kept out of the index23
Agentic work benchmarks for GPT-5 and GPT-5.2
BenchmarkGPT-5GPT-5.2Leader
Design Arena: agents building web appsNot measured1086Only GPT-5.2 measured
Design Arena: agents building full-stack appsNot measured1043Only GPT-5.2 measured
Design Arena: agents building mobile appsNot measured1094Only GPT-5.2 measured
Design Arena: agents building Android appsNot measured1064Only GPT-5.2 measured
Design Arena: agents building Godot gamesNot measured1142Only GPT-5.2 measured
τ²-Bench Telecom86.5%84.8%GPT-5
METR time horizons69.6%75.3%GPT-5.2
Remote Labor Index1.7%2.5%GPT-5.2
GDPval34.8%49.7%GPT-5.2
BALROG32.8%Not measuredOnly GPT-5 measured
DeepResearch Bench49.6%41.1%GPT-5
Vending-Bench 2Not measured$3,591Only GPT-5.2 measured
Neither model has a result on the other 27 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index214
Reasoning benchmarks for GPT-5 and GPT-5.2
BenchmarkGPT-5GPT-5.2Leader
GPQA Diamond85.6%91.7%GPT-5.2
MMMU Pro (Vals AI)81.5%86.7%GPT-5.2
GPQA Diamond (Artificial Analysis run)85.4%90.3%GPT-5.2
Humanity's Last Exam (Artificial Analysis run)28.5%37.7%GPT-5.2
GPQA Diamond (Epoch AI run)86.2%91.4%GPT-5.2
ARC-AGI65.7%86.2%GPT-5.2
ARC-AGI-29.9%52.9%GPT-5.2
SimpleBench56.7%45.8%GPT-5
Chess puzzles36.0%48.0%GPT-5.2
Mystery game puzzles23.0%23.0%Tie
EBR-bench12.7%23.0%GPT-5.2
VPCT66.0%84.0%GPT-5.2
LMArena Text: hard prompts14151428GPT-5.2
LMArena Text: expert14191445GPT-5.2
LMArena Text: hard prompts in English14191431GPT-5.2
LMArena Vision: diagrams12521265GPT-5.2
LMArena Vision: homework12781276GPT-5
Neither model has a result on the other 4 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index19
Math benchmarks for GPT-5 and GPT-5.2
BenchmarkGPT-5GPT-5.2Leader
AIME93.4%96.9%GPT-5.2
MATH 500 (Vals AI)96.0%Not measuredOnly GPT-5 measured
MGSM (Vals AI)92.8%94.0%GPT-5.2
AA Math Index94.399.0GPT-5.2
AIME 2025 (Artificial Analysis run)94.3%99.0%GPT-5.2
OTIS Mock AIME91.4%96.1%GPT-5.2
FrontierMath, tiers 1 to 355.4%67.4%GPT-5.2
FrontierMath, tier 422.0%31.7%GPT-5.2
ProofBench18.0%15.0%GPT-5
MATH level 598.1%Not measuredOnly GPT-5 measured
LMArena Text: math14071440GPT-5.2
LMArena Text: mathematical work14051442GPT-5.2
Neither model has a result on the other 1 board here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index68
Knowledge benchmarks for GPT-5 and GPT-5.2
BenchmarkGPT-5GPT-5.2Leader
MMLU Pro86.5%86.2%GPT-5
LegalBench (Vals AI)86.0%82.8%GPT-5
CaseLaw (Vals AI)66.5%66.0%GPT-5
CorpFin (Vals AI)61.1%65.9%GPT-5.2
TaxEval (Vals AI)73.4%75.8%GPT-5.2
MortgageTax (Vals AI)65.5%67.1%GPT-5.2
MedQA (Vals AI)96.3%94.1%GPT-5
MedCode (Vals AI)49.6%49.8%GPT-5.2
MMLU Pro (Artificial Analysis run)87.1%87.4%GPT-5.2
SimpleQA Verified50.1%37.1%GPT-5
GeoBench81.0%Not measuredOnly GPT-5 measured
LMArena Text: science14161425GPT-5.2
LMArena Text: medicine and healthcare14131425GPT-5.2
LMArena Text: legal and government14311427GPT-5
LMArena Text: business and finance14001414GPT-5.2
Neither model has a result on the other 2 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index26
Web design benchmarks for GPT-5 and GPT-5.2
BenchmarkGPT-5GPT-5.2Leader
Design Arena: websites11851196GPT-5.2
Design Arena: UI components11791190GPT-5.2
Design Arena: game dev11851196GPT-5.2
Design Arena: data viz12211202GPT-5
Design Arena: 3D10561079GPT-5.2
Design Arena: all code categories11681171GPT-5.2
Design Arena: SVG11881137GPT-5
Design Arena: ASCII art11401153GPT-5.2
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index427
General benchmarks for GPT-5 and GPT-5.2
BenchmarkGPT-5GPT-5.2Leader
SAGE (Vals AI)43.7%49.3%GPT-5.2
MedScribe (Vals AI)83.7%84.4%GPT-5.2
IFBench73.1%75.4%GPT-5.2
Long context reasoning (AA-LCR)78.2%82.7%GPT-5.2
DTBench90.7%90.9%GPT-5.2
LMCA40.0%43.9%GPT-5.2
Fiction.LiveBench97.2%Not measuredOnly GPT-5 measured
Creative writing (Lech Mazur)86.0%Not measuredOnly GPT-5 measured
CL-benchNot measured18.2%Only GPT-5.2 measured
Furniture assemblyNot measured38.3%Only GPT-5.2 measured
LMArena Text14061416GPT-5.2
LMArena Text: instruction following13881409GPT-5.2
LMArena Text: creative writing13651376GPT-5.2
LMArena Vision12321246GPT-5.2
LMArena Text: English14091422GPT-5.2
LMArena Text: non-English13971405GPT-5.2
LMArena Text: Chinese14221447GPT-5.2
LMArena Text: German14161444GPT-5.2
LMArena Text: Japanese14091420GPT-5.2
LMArena Text: Korean13601374GPT-5.2
LMArena Text: French14091454GPT-5.2
LMArena Text: Spanish13991413GPT-5.2
LMArena Text: Russian14071415GPT-5.2
LMArena Text: Polish14311426GPT-5
LMArena Text: multi-turn14261419GPT-5
LMArena Text: longer queries13991413GPT-5.2
LMArena Text: writing and language13841392GPT-5.2
LMArena Text: entertainment and media13791378GPT-5
LMArena Vision: English12311243GPT-5.2
LMArena Vision: Chinese12671305GPT-5.2
LMArena Vision: reading text12461257GPT-5.2
LMArena Vision: humor12191227GPT-5.2
LMArena Vision: captioning12121267GPT-5.2
LMArena Vision: recognizing things12631232GPT-5
LMArena Vision: creative writing12101225GPT-5.2
LMArena DocumentNot measured1405Only GPT-5.2 measured
Neither model has a result on the other 10 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-5.2
BenchmarkGPT-5GPT-5.2Leader
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-5.2$0.0088Fits in one request

GPT-5 costs less

Repository review

50K input + 3K output tokens

GPT-5$0.092Fits in one request
GPT-5.2$0.130Fits in one request

GPT-5 costs less

Agent loop

200K input + 10K output tokens

GPT-5$0.350Fits in one request
GPT-5.2$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-5.2
SpecificationOpenAI GPT-5OpenAI GPT-5.2
MakerOpenAIOpenAI
WeightsNo public weights listedNo public weights listed
ReleasedAugust 7, 2025December 10, 2025
Context window400K400K
Input $/Mtok$1.25$1.75
Output $/Mtok$10.00$14.00
Coding index72.258.9
Independent measurements behind it32
Cost per solved task$0.39$1.91
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-5 better than GPT-5.2 for coding?

Pick GPT-5 for coding: it scores 72.2 on the Vibe Coding Index against 58.9 for GPT-5.2. Pick GPT-5.2 when what matters is SWE-bench Verified (Vals AI) or Vibe Code Bench (Vals AI), where it leads.

Which is cheaper, GPT-5 or GPT-5.2?

GPT-5 costs $1.25 per million input tokens and $10.00 per million output tokens; GPT-5.2 costs $1.75 and $14.00. Per solved SWE-bench Verified task, GPT-5 costs $0.39 and GPT-5.2 $1.91.

Which has the bigger context window, GPT-5 or GPT-5.2?

Neither: both take 400K tokens.

How many benchmarks is the GPT-5 vs GPT-5.2 comparison based on?

101 benchmarks have a result for both models as of October 21, 2026. GPT-5's coding index rests on 3 independent benchmarks and GPT-5.2's on 2. 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-5-2" title="GPT-5 vs GPT-5.2, 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-5-2">LLMBenchmark.ai, GPT-5 vs GPT-5.2</a></p>