GPT-5 vs o3

GPT-5 leads o3 on the Artificial Analysis Intelligence Index, 23.0 to 20.2, and o3 costs 1.3x less per output token ($8.00 vs $10.00 per million). Updated October 21, 2026; 98 benchmarks have a result for both.

Replace GPT-5 Replace o3

GPT-5OpenAI

Coding: 38.8
Overall: 51.5
Output $/M: $10.00

o3OpenAI

Coding: No data
Overall: 51.1
Output $/M: $8.00

vs boards won of 98 shared

GPT-5 or o3: which to pick

Only GPT-5 has a coding score, 38.8 on the Vibe Coding Index; o3 has none, so the two cannot be ranked on coding.

Pick o3 for output tokens 1.3x cheaper ($8.00 vs $10.00 per million) and work measured by GSO-Bench or ALE-Bench, where it leads.

Model AGPT-5OpenAI38.8/1003 independent benchmarks$0.39 per solved task Model Bo3OpenAINot benchmarkedNo coding benchmark$0.571 per solved task

98154

Shared results: 98
GPT-5 only: 15
o3 only: 4
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.

Coding
Vibe Coding Index, six public coding benchmarks
No call
Only GPT-5 has a coding score, so no call is made.
One side not measured
Intelligence
AA Intelligence Index
GPT-5
GPT-5 leads, 23.0 against 20.2.
Same board, both measured
Reasoning
GPQA Diamond
GPT-5
GPT-5 leads, 85.6% against 84.1%.
Same board, both measured
Math
AIME
GPT-5
GPT-5 leads, 93.4% against 85.3%.
Same board, both measured
Knowledge
MMLU Pro
GPT-5
GPT-5 leads, 86.5% against 85.6%.
Same board, both measured
Web design
Design Arena: websites
GPT-5
GPT-5 leads, 1184 against 1036.
Same board, both measured
Long documents
Listed context window
GPT-5
GPT-5 takes 400K tokens against 200K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
o3
GPT-5 $0.0063, o3 $0.0060 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
GPT-5
GPT-5 $0.39, o3 $0.571.
Published attempt cost and pass rate

At a glance

Coding index: GPT-5 38.8 from 3 benchmarks, o3 not benchmarked.
Widest category: Web design, 15.8 points apart on average over 5 shared benchmarks.
Widest single gap: FrontierMath, tiers 1 to 3: GPT-5 55.4% against o3 33.3%, 22.1 points apart on a 0 to 100 scale.
Output price: GPT-5 $10.00 against o3 $8.00 per million output tokens, 1.3x apart.
Context window: GPT-5 takes 400K tokens, o3 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
Category	GPT-5	o3
Coding	8	3
Intelligence	2	0
Agentic work	5	0
Reasoning	15	2
Math	8	2
Knowledge	11	4
Web design	5	0
General	24	9
All shared boards	78	20

CodingLike-for-likeSix of these boards feed the Vibe Coding Index83

Coding benchmarks for GPT-5 and o3
Benchmark	GPT-5	o3	Leader
Aider polyglot	88.0%	81.3%	GPT-5
SWE-bench Verified	74.4%	58.4%	GPT-5
SWE-bench Multimodal	Not measured	36.0%	Only o3 measured
LMArena WebDev	1418	Not measured	Only GPT-5 measured
AA Coding Index	37.8	Not measured	Only GPT-5 measured
LiveCodeBench (Vals AI)	85.9%	83.9%	GPT-5
SWE-bench Verified (Vals AI)	69.0%	Not measured	Only GPT-5 measured
Vibe Code Bench (Vals AI)	20.1%	Not measured	Only GPT-5 measured
LiveCodeBench (Artificial Analysis run)	84.6%	80.8%	GPT-5
Terminal-Bench Hard (Artificial Analysis run)	37.9%	37.1%	GPT-5
Terminal-Bench	49.6%	Not measured	Only GPT-5 measured
GSO-Bench	6.9%	8.8%	o3
WeirdML	60.7%	52.4%	GPT-5
ALE-Bench	807.6	933.5	o3
LMArena Text: coding	1436	1407	GPT-5
LMArena Text: software and IT	1420	1415	GPT-5
LMArena WebDev: HTML	1427	Not measured	Only GPT-5 measured
SciCode	0.0%	9.2%	o3
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 o3
Benchmark	GPT-5	o3	Leader
AA Intelligence Index	23.0	20.2	GPT-5
Epoch Capabilities Index	150.0	146.7	GPT-5

Agentic workLike-for-likePublic boards, shown as published, kept out of the index50

Agentic work benchmarks for GPT-5 and o3
Benchmark	GPT-5	o3	Leader
τ²-Bench Telecom	86.5%	80.7%	GPT-5
METR time horizons	69.6%	65.4%	GPT-5
Remote Labor Index	1.7%	Not measured	Only GPT-5 measured
GDPval	34.8%	30.8%	GPT-5
BALROG	32.8%	Not measured	Only GPT-5 measured
DeepResearch Bench	49.6%	45.2%	GPT-5
OSWorld	Not measured	23.0%	Only o3 measured
Berkeley Function Calling Leaderboard (BFCL)	Not measured	63.0%	Only o3 measured
BrowseComp-Plus	80.0%	65.9%	GPT-5
Neither model has a result on the other 32 boards here

ReasoningLike-for-likePublic boards, shown as published, kept out of the index152

Reasoning benchmarks for GPT-5 and o3
Benchmark	GPT-5	o3	Leader
GPQA Diamond	85.6%	84.1%	GPT-5
MMMU Pro (Vals AI)	81.5%	80.4%	GPT-5
GPQA Diamond (Artificial Analysis run)	85.4%	82.7%	GPT-5
Humanity's Last Exam (Artificial Analysis run)	28.5%	20.1%	GPT-5
GPQA Diamond (Epoch AI run)	86.2%	81.8%	GPT-5
ARC-AGI	65.7%	60.8%	GPT-5
ARC-AGI-2	9.9%	6.5%	GPT-5
SimpleBench	56.7%	53.1%	GPT-5
Chess puzzles	36.0%	37.0%	o3
Mystery game puzzles	23.0%	29.0%	o3
EBR-bench	12.7%	Not measured	Only GPT-5 measured
VPCT	66.0%	52.0%	GPT-5
LMArena Text: hard prompts	1415	1402	GPT-5
LMArena Text: expert	1419	1402	GPT-5
LMArena Text: hard prompts in English	1419	1411	GPT-5
LMArena Vision: diagrams	1252	1220	GPT-5
LMArena Vision: homework	1278	1234	GPT-5
Kagi LLM Benchmark	72.7%	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 index82

Math benchmarks for GPT-5 and o3
Benchmark	GPT-5	o3	Leader
AIME	93.4%	85.3%	GPT-5
MATH 500 (Vals AI)	96.0%	94.6%	GPT-5
MGSM (Vals AI)	92.8%	91.8%	GPT-5
AA Math Index	94.3	88.3	GPT-5
AIME 2025 (Artificial Analysis run)	94.3%	88.3%	GPT-5
OTIS Mock AIME	91.4%	84.4%	GPT-5
FrontierMath, tiers 1 to 3	55.4%	33.3%	GPT-5
FrontierMath, tier 4	22.0%	Not measured	Only GPT-5 measured
ProofBench	18.0%	Not measured	Only GPT-5 measured
MATH level 5	98.1%	97.8%	GPT-5
LMArena Text: math	1407	1426	o3
LMArena Text: mathematical work	1405	1423	o3
Neither model has a result on the other 1 board here

KnowledgeLike-for-likePublic boards, shown as published, kept out of the index114

Knowledge benchmarks for GPT-5 and o3
Benchmark	GPT-5	o3	Leader
MMLU Pro	86.5%	85.6%	GPT-5
LegalBench (Vals AI)	86.0%	83.8%	GPT-5
CaseLaw (Vals AI)	66.5%	Not measured	Only GPT-5 measured
CorpFin (Vals AI)	61.1%	59.7%	GPT-5
TaxEval (Vals AI)	73.4%	74.6%	o3
MortgageTax (Vals AI)	65.5%	65.7%	o3
MedQA (Vals AI)	96.3%	96.1%	GPT-5
MedCode (Vals AI)	49.6%	47.3%	GPT-5
MMLU Pro (Artificial Analysis run)	87.1%	85.3%	GPT-5
SimpleQA Verified	50.1%	49.4%	GPT-5
GeoBench	81.0%	74.0%	GPT-5
LMArena Text: science	1416	1429	o3
LMArena Text: medicine and healthcare	1413	1448	o3
LMArena Text: legal and government	1431	1428	GPT-5
LMArena Text: business and finance	1400	1393	GPT-5
FACTS Benchmark Suite	56.5%	44.4%	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 index50

Web design benchmarks for GPT-5 and o3
Benchmark	GPT-5	o3	Leader
Design Arena: websites	1184	1036	GPT-5
Design Arena: UI components	1180	1017	GPT-5
Design Arena: game dev	1185	1034	GPT-5
Design Arena: data viz	1221	1200	GPT-5
Design Arena: 3D	1056	Not measured	Only GPT-5 measured
Design Arena: all code categories	1168	1020	GPT-5
Design Arena: SVG	1189	Not measured	Only GPT-5 measured
Design Arena: ASCII art	1141	Not measured	Only 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 index249

General benchmarks for GPT-5 and o3
Benchmark	GPT-5	o3	Leader
SAGE (Vals AI)	43.7%	41.8%	GPT-5
MedScribe (Vals AI)	83.7%	76.7%	GPT-5
IFBench	73.1%	71.4%	GPT-5
Long context reasoning (AA-LCR)	78.2%	74.7%	GPT-5
DTBench	90.7%	84.8%	GPT-5
LMCA	40.0%	39.7%	GPT-5
Fiction.LiveBench	97.2%	88.9%	GPT-5
Creative writing (Lech Mazur)	86.0%	83.9%	GPT-5
CL-bench	Not measured	17.8%	Only o3 measured
LMArena Text	1406	1410	o3
LMArena Text: instruction following	1388	1368	GPT-5
LMArena Text: creative writing	1365	1359	GPT-5
LMArena Vision	1232	1214	GPT-5
LMArena Text: English	1409	1417	o3
LMArena Text: non-English	1397	1401	o3
LMArena Text: Chinese	1422	1437	o3
LMArena Text: German	1416	1420	o3
LMArena Text: Japanese	1409	1403	GPT-5
LMArena Text: Korean	1360	1370	o3
LMArena Text: French	1409	1430	o3
LMArena Text: Spanish	1399	1395	GPT-5
LMArena Text: Russian	1407	1406	GPT-5
LMArena Text: Polish	1431	1438	o3
LMArena Text: multi-turn	1426	1405	GPT-5
LMArena Text: longer queries	1399	1372	GPT-5
LMArena Text: writing and language	1384	1372	GPT-5
LMArena Text: entertainment and media	1379	1373	GPT-5
LMArena Vision: English	1231	1221	GPT-5
LMArena Vision: Chinese	1267	1237	GPT-5
LMArena Vision: reading text	1246	1218	GPT-5
LMArena Vision: humor	1219	1222	o3
LMArena Vision: captioning	1212	1210	GPT-5
LMArena Vision: recognizing things	1263	1234	GPT-5
LMArena Vision: creative writing	1210	1201	GPT-5
Neither model has a result on the other 12 boards here

Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared

Image generation benchmarks for GPT-5 and o3
Benchmark	GPT-5	o3	Leader
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

o3$0.0060Fits in one request

o3 costs less

Repository review

50K input + 3K output tokens

GPT-5$0.092Fits in one request

o3$0.124Fits in one request

GPT-5 costs less

Agent loop

200K input + 10K output tokens

GPT-5$0.350Fits in one request

o3$0.480Exceeds the context window

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 o3
Specification	OpenAI GPT-5	OpenAI o3
Maker	OpenAI	OpenAI
Weights	No public weights listed	No public weights listed
Released	August 7, 2025	April 16, 2025
Context window	400K	200K
Input $/Mtok	$1.25	$2.00
Output $/Mtok	$10.00	$8.00
Coding index	38.8	Not benchmarked
Independent measurements behind it	3	0
Cost per solved task	$0.39	$0.571
Source	OpenRouter listing	OpenRouter listing

Questions

Is GPT-5 better than o3 for coding?

Only GPT-5 has a coding score, 38.8 on the Vibe Coding Index; o3 has none, so the two cannot be ranked on coding. Pick o3 for output tokens 1.3x cheaper ($8.00 vs $10.00 per million) and work measured by GSO-Bench or ALE-Bench, where it leads.

Which is cheaper, GPT-5 or o3?

GPT-5 costs $1.25 per million input tokens and $10.00 per million output tokens; o3 costs $2.00 and $8.00. Per solved SWE-bench Verified task, GPT-5 costs $0.39 and o3 $0.571.

Which has the bigger context window, GPT-5 or o3?

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

How many benchmarks is the GPT-5 vs o3 comparison based on?

98 benchmarks have a result for both models as of October 21, 2026. GPT-5's coding index rests on 3 independent benchmarks and o3's on 0. Where only one model has a result, the page says so instead of naming a winner.

GPT-5 or o3: which to pick

Which one for your work

Coding

Intelligence

Reasoning

Math

Knowledge

Web design

Long documents

Everyday cost

Cost per solved task