GPT-4.1 vs GPT-5

Updated October 21, 2026. 86 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace GPT-4.1 Replace GPT-5

GPT-4.1OpenAI

Coding: 46.5
Overall: 12.7
Output $/M: $8.00

GPT-5OpenAI

Coding: 72.2
Overall: 23.0
Output $/M: $10.00

vs boards won of 86 shared

Decision reading

Pick GPT-5 for coding: it scores 72.2 on the Vibe Coding Index against 46.5 for GPT-4.1.

Pick GPT-4.1 when output tokens cost 1.3x less, or when a solved task costs $0.37 against $0.39.

Model AGPT-4.1OpenAI46.5/1002 independent benchmarks$0.37 per solved task Model BGPT-5OpenAI72.2/1003 independent benchmarks$0.39 per solved task

86123

Shared results: 86
GPT-4.1 only: 1
GPT-5 only: 23
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.

Coding
Vibe Coding Index, four public coding benchmarks
GPT-5
GPT-5 scores 72.2 against 46.5.
2 and 3 benchmarks behind the scores
Intelligence
AA Intelligence Index
GPT-5
GPT-5 leads, 23.0 against 12.7.
Same board, both measured
Reasoning
GPQA Diamond
GPT-5
GPT-5 leads, 85.6% against 65.4%.
Same board, both measured
Math
AIME
GPT-5
GPT-5 leads, 93.4% against 39.6%.
Same board, both measured
Knowledge
MMLU Pro
GPT-5
GPT-5 leads, 86.5% against 80.5%.
Same board, both measured
Web design
Design Arena: websites
GPT-5
GPT-5 leads, 1185 against 1039.
Same board, both measured
Long documents
Listed context window
GPT-4.1
GPT-4.1 takes 1.05M tokens against 400K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
GPT-4.1
GPT-4.1 $0.0060, GPT-5 $0.0063 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
GPT-4.1
GPT-4.1 $0.37, GPT-5 $0.39.
Published attempt cost and pass rate

At a glance

Coding index: GPT-4.1 46.5 from 2 benchmarks, GPT-5 72.2 from 3 benchmarks.
Widest category: Agentic work, 39.5 points apart on average over 1 shared benchmark.
Widest single gap: ARC-AGI: GPT-4.1 5.5% against GPT-5 65.7%, 60.2 points apart on a 0 to 100 scale.
Output price: GPT-4.1 $8.00 against GPT-5 $10.00 per million output tokens, 1.3x apart.
Context window: GPT-4.1 takes 1.05M tokens, GPT-5 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
Category	GPT-4.1	GPT-5
Coding	0	9
Intelligence	0	2
Agentic work	0	1
Reasoning	0	14
Math	0	10
Knowledge	4	10
Web design	0	6
General	0	30
All shared boards	4	82

CodingLike-for-likeFour of these boards feed the Vibe Coding Index09

Coding benchmarks for GPT-4.1 and GPT-5
Benchmark	GPT-4.1	GPT-5	Leader
Aider polyglot	52.4%	88.0%	GPT-5
SWE-bench Verified	39.6%	74.4%	GPT-5
SWE-bench Multimodal	31.1%	Not measured	Only GPT-4.1 measured
LMArena WebDev	Not measured	1418	Only GPT-5 measured
AA Coding Index	Not measured	37.8	Only GPT-5 measured
LiveCodeBench (Vals AI)	54.7%	85.9%	GPT-5
SWE-bench Verified (Vals AI)	Not measured	69.0%	Only GPT-5 measured
Vibe Code Bench (Vals AI)	Not measured	20.1%	Only GPT-5 measured
LiveCodeBench (Artificial Analysis run)	45.7%	84.6%	GPT-5
Terminal-Bench Hard (Artificial Analysis run)	13.6%	37.9%	GPT-5
Terminal-Bench	Not measured	49.6%	Only GPT-5 measured
GSO-Bench	Not measured	6.9%	Only GPT-5 measured
WeirdML	39.0%	60.7%	GPT-5
ALE-Bench	558.1	807.6	GPT-5
LMArena Text: coding	1390	1436	GPT-5
LMArena Text: software and IT	1391	1420	GPT-5
LMArena WebDev: HTML	Not measured	1427	Only GPT-5 measured
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-4.1 and GPT-5
Benchmark	GPT-4.1	GPT-5	Leader
AA Intelligence Index	12.7	23.0	GPT-5
Epoch Capabilities Index	136.7	150.0	GPT-5

Agentic workLike-for-likePublic boards, shown as published, kept out of the index01

Agentic work benchmarks for GPT-4.1 and GPT-5
Benchmark	GPT-4.1	GPT-5	Leader
τ²-Bench Telecom	47.1%	86.5%	GPT-5
METR time horizons	Not measured	69.6%	Only GPT-5 measured
Remote Labor Index	Not measured	1.7%	Only GPT-5 measured
GDPval	Not measured	34.8%	Only GPT-5 measured
BALROG	Not measured	32.8%	Only GPT-5 measured
DeepResearch Bench	Not measured	49.6%	Only 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 index014

Reasoning benchmarks for GPT-4.1 and GPT-5
Benchmark	GPT-4.1	GPT-5	Leader
GPQA Diamond	65.4%	85.6%	GPT-5
MMMU Pro (Vals AI)	72.4%	81.5%	GPT-5
GPQA Diamond (Artificial Analysis run)	66.6%	85.4%	GPT-5
Humanity's Last Exam (Artificial Analysis run)	4.2%	28.5%	GPT-5
GPQA Diamond (Epoch AI run)	66.9%	86.2%	GPT-5
ARC-AGI	5.5%	65.7%	GPT-5
ARC-AGI-2	0.4%	9.9%	GPT-5
SimpleBench	27.0%	56.7%	GPT-5
Chess puzzles	3.0%	36.0%	GPT-5
Mystery game puzzles	Not measured	23.0%	Only GPT-5 measured
EBR-bench	Not measured	12.7%	Only GPT-5 measured
VPCT	Not measured	66.0%	Only GPT-5 measured
LMArena Text: hard prompts	1384	1415	GPT-5
LMArena Text: expert	1364	1419	GPT-5
LMArena Text: hard prompts in English	1393	1419	GPT-5
LMArena Vision: diagrams	1215	1252	GPT-5
LMArena Vision: homework	1248	1278	GPT-5
Neither model has a result on the other 4 boards here

MathLike-for-likePublic boards, shown as published, kept out of the index010

Math benchmarks for GPT-4.1 and GPT-5
Benchmark	GPT-4.1	GPT-5	Leader
AIME	39.6%	93.4%	GPT-5
MATH 500 (Vals AI)	87.2%	96.0%	GPT-5
MGSM (Vals AI)	87.7%	92.8%	GPT-5
AA Math Index	34.7	94.3	GPT-5
AIME 2025 (Artificial Analysis run)	34.7%	94.3%	GPT-5
OTIS Mock AIME	38.3%	91.4%	GPT-5
FrontierMath, tiers 1 to 3	6.0%	55.4%	GPT-5
FrontierMath, tier 4	Not measured	22.0%	Only GPT-5 measured
ProofBench	Not measured	18.0%	Only GPT-5 measured
MATH level 5	83.0%	98.1%	GPT-5
LMArena Text: math	1369	1407	GPT-5
LMArena Text: mathematical work	1360	1405	GPT-5
Neither model has a result on the other 1 board here

KnowledgeLike-for-likePublic boards, shown as published, kept out of the index410

Knowledge benchmarks for GPT-4.1 and GPT-5
Benchmark	GPT-4.1	GPT-5	Leader
MMLU Pro	80.5%	86.5%	GPT-5
LegalBench (Vals AI)	83.1%	86.0%	GPT-5
CaseLaw (Vals AI)	69.9%	66.5%	GPT-4.1
CorpFin (Vals AI)	63.0%	61.1%	GPT-4.1
TaxEval (Vals AI)	75.1%	73.4%	GPT-4.1
MortgageTax (Vals AI)	65.9%	65.5%	GPT-4.1
MedQA (Vals AI)	91.2%	96.3%	GPT-5
MedCode (Vals AI)	Not measured	49.6%	Only GPT-5 measured
MMLU Pro (Artificial Analysis run)	80.6%	87.1%	GPT-5
SimpleQA Verified	31.1%	50.1%	GPT-5
GeoBench	72.0%	81.0%	GPT-5
LMArena Text: science	1382	1416	GPT-5
LMArena Text: medicine and healthcare	1379	1413	GPT-5
LMArena Text: legal and government	1399	1431	GPT-5
LMArena Text: business and finance	1369	1400	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 index06

Web design benchmarks for GPT-4.1 and GPT-5
Benchmark	GPT-4.1	GPT-5	Leader
Design Arena: websites	1039	1185	GPT-5
Design Arena: UI components	1000	1179	GPT-5
Design Arena: game dev	1080	1185	GPT-5
Design Arena: data viz	1104	1221	GPT-5
Design Arena: 3D	850	1056	GPT-5
Design Arena: all code categories	1026	1168	GPT-5
Design Arena: SVG	Not measured	1188	Only GPT-5 measured
Design Arena: ASCII art	Not measured	1140	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 index030

General benchmarks for GPT-4.1 and GPT-5
Benchmark	GPT-4.1	GPT-5	Leader
SAGE (Vals AI)	Not measured	43.7%	Only GPT-5 measured
MedScribe (Vals AI)	Not measured	83.7%	Only GPT-5 measured
IFBench	43.0%	73.1%	GPT-5
Long context reasoning (AA-LCR)	68.3%	78.2%	GPT-5
DTBench	68.3%	90.7%	GPT-5
LMCA	25.6%	40.0%	GPT-5
Fiction.LiveBench	63.9%	97.2%	GPT-5
Creative writing (Lech Mazur)	Not measured	86.0%	Only GPT-5 measured
LMArena Text	1383	1406	GPT-5
LMArena Text: instruction following	1367	1388	GPT-5
LMArena Text: creative writing	1363	1365	GPT-5
LMArena Vision	1210	1232	GPT-5
LMArena Text: English	1392	1409	GPT-5
LMArena Text: non-English	1370	1397	GPT-5
LMArena Text: Chinese	1382	1422	GPT-5
LMArena Text: German	1381	1416	GPT-5
LMArena Text: Japanese	1319	1409	GPT-5
LMArena Text: Korean	1338	1360	GPT-5
LMArena Text: French	1382	1409	GPT-5
LMArena Text: Spanish	1376	1399	GPT-5
LMArena Text: Russian	1377	1407	GPT-5
LMArena Text: Polish	1389	1431	GPT-5
LMArena Text: multi-turn	1397	1426	GPT-5
LMArena Text: longer queries	1385	1399	GPT-5
LMArena Text: writing and language	1371	1384	GPT-5
LMArena Text: entertainment and media	1365	1379	GPT-5
LMArena Vision: English	1209	1231	GPT-5
LMArena Vision: Chinese	1226	1267	GPT-5
LMArena Vision: reading text	1218	1246	GPT-5
LMArena Vision: humor	1196	1219	GPT-5
LMArena Vision: captioning	1207	1212	GPT-5
LMArena Vision: recognizing things	1208	1263	GPT-5
LMArena Vision: creative writing	1210	1210	GPT-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-4.1 and GPT-5
Benchmark	GPT-4.1	GPT-5	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-4.1$0.0060Fits in one request

GPT-5$0.0063Fits in one request

GPT-4.1 costs less

Repository review

50K input + 3K output tokens

GPT-4.1$0.124Fits in one request

GPT-5$0.092Fits in one request

GPT-5 costs less

Agent loop

200K input + 10K output tokens

GPT-4.1$0.480Fits in one request

GPT-5$0.350Fits 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-4.1 and GPT-5
Specification	OpenAI GPT-4.1	OpenAI GPT-5
Maker	OpenAI	OpenAI
Weights	No public weights listed	No public weights listed
Released	April 14, 2025	August 7, 2025
Context window	1.05M	400K
Input $/Mtok	$2.00	$1.25
Output $/Mtok	$8.00	$10.00
Coding index	46.5	72.2
Independent measurements behind it	2	3
Cost per solved task	$0.37	$0.39
Source	OpenRouter listing	OpenRouter listing

Questions

Is GPT-4.1 better than GPT-5 for coding?

Pick GPT-5 for coding: it scores 72.2 on the Vibe Coding Index against 46.5 for GPT-4.1. Pick GPT-4.1 when output tokens cost 1.3x less, or when a solved task costs $0.37 against $0.39.

Which is cheaper, GPT-4.1 or GPT-5?

GPT-4.1 costs $2.00 per million input tokens and $8.00 per million output tokens; GPT-5 costs $1.25 and $10.00. Per solved SWE-bench Verified task, GPT-4.1 costs $0.37 and GPT-5 $0.39.

Which has the bigger context window, GPT-4.1 or GPT-5?

GPT-4.1, with 1.05M tokens against 400K.

How many benchmarks is the GPT-4.1 vs GPT-5 comparison based on?

86 benchmarks have a result for both models as of October 21, 2026. GPT-4.1's coding index rests on 2 independent benchmarks and GPT-5'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-4-1-vs-gpt-5" title="GPT-4.1 vs GPT-5, 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-4-1-vs-gpt-5">LLMBenchmark.ai, GPT-4.1 vs GPT-5</a></p>

Decision reading

Which one for your work

Coding

Intelligence

Reasoning

Math

Knowledge

Web design

Long documents

Everyday cost

Cost per solved task