GPT-4.1 vs GPT-5.2-Codex
Updated October 21, 2026. 10 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.
Decision reading
Pick GPT-5.2-Codex for coding: it scores 55.1 on the Vibe Coding Index against 46.5 for GPT-4.1.
Pick GPT-4.1 when output tokens cost 1.8x less, or when a solved task costs $0.37 against $0.43.
- Shared results
- 10
- GPT-4.1 only
- 77
- GPT-5.2-Codex only
- 12
- Like-for-like categories
- 5 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-5.2-Codex
GPT-5.2-Codex scores 55.1 against 46.5.
2 benchmarks behind each score
Intelligence
AA Intelligence Index
GPT-5.2-Codex
GPT-5.2-Codex leads, 28.5 against 12.7.
Same board, both measured
Reasoning
GPQA Diamond
No call
Only GPT-4.1 has a GPQA Diamond result, 65.4%.
Directional only
Math
AIME
No call
Only GPT-4.1 has a AIME result, 39.6%.
Directional only
Knowledge
MMLU Pro
No call
Only GPT-4.1 has a MMLU Pro result, 80.5%.
Directional only
Web design
Design Arena: websites
No call
Only GPT-4.1 has a Design Arena: websites result, 1039.
Directional only
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-5.2-Codex $0.0088 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
GPT-4.1
GPT-4.1 $0.37, GPT-5.2-Codex $0.43.
Published attempt cost and pass rate
At a glance
- Coding index
- GPT-4.1 46.5 from 2 benchmarks, GPT-5.2-Codex 55.1 from 2 benchmarks.
- Widest category
- Agentic work, 45.0 points apart on average over 1 shared benchmark.
- Widest single gap
- τ²-Bench Telecom: GPT-4.1 47.1% against GPT-5.2-Codex 92.1%, 45.0 points apart on a 0 to 100 scale.
- Output price
- GPT-4.1 $8.00 against GPT-5.2-Codex $14.00 per million output tokens, 1.8x apart.
- Context window
- GPT-4.1 takes 1.05M tokens, GPT-5.2-Codex 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.
| Category | GPT-4.1 | Bars | GPT-5.2-Codex | Ties |
|---|---|---|---|---|
| Coding | 0 | 4 | ||
| Intelligence | 0 | 1 | ||
| Agentic work | 0 | 1 | ||
| Reasoning | 0 | 2 | ||
| General | 0 | 2 | ||
| All shared boards | 0 | 10 |
CodingLike-for-likeFour of these boards feed the Vibe Coding Index04
| Benchmark | GPT-4.1 | GPT-5.2-Codex | Leader |
|---|---|---|---|
| Aider polyglot | 52.4% | Not measured | Only GPT-4.1 measured |
| SWE-bench Verified | 39.6% | 72.8% | GPT-5.2-Codex |
| SWE-bench Multimodal | 31.1% | Not measured | Only GPT-4.1 measured |
| SWE-bench Multilingual | Not measured | 66.3% | Only GPT-5.2-Codex measured |
| LMArena WebDev | Not measured | 1339 | Only GPT-5.2-Codex measured |
| LiveCodeBench (Vals AI) | 54.7% | 88.0% | GPT-5.2-Codex |
| SWE-bench Verified (Vals AI) | Not measured | 72.4% | Only GPT-5.2-Codex measured |
| Vibe Code Bench (Vals AI) | Not measured | 37.9% | Only GPT-5.2-Codex measured |
| LiveCodeBench (Artificial Analysis run) | 45.7% | Not measured | Only GPT-4.1 measured |
| Terminal-Bench Hard (Artificial Analysis run) | 13.6% | 37.1% | GPT-5.2-Codex |
| Terminal-Bench | Not measured | 66.5% | Only GPT-5.2-Codex measured |
| WeirdML | 39.0% | Not measured | Only GPT-4.1 measured |
| ALE-Bench | 558.1 | 1299.9 | GPT-5.2-Codex |
| LMArena Text: coding | 1390 | Not measured | Only GPT-4.1 measured |
| LMArena Text: software and IT | 1391 | Not measured | Only GPT-4.1 measured |
| LMArena WebDev: React | Not measured | 1329 | Only GPT-5.2-Codex measured |
| LMArena WebDev: HTML | Not measured | 1348 | Only GPT-5.2-Codex measured |
| Neither model has a result on the other 21 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | GPT-4.1 | GPT-5.2-Codex | Leader |
|---|---|---|---|
| AA Intelligence Index | 12.7 | 28.5 | GPT-5.2-Codex |
| Epoch Capabilities Index | 136.7 | Not measured | Only GPT-4.1 measured |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | GPT-4.1 | GPT-5.2-Codex | Leader |
|---|---|---|---|
| Design Arena: agents building web apps | Not measured | 1051 | Only GPT-5.2-Codex measured |
| Design Arena: agents building full-stack apps | Not measured | 993 | Only GPT-5.2-Codex measured |
| Design Arena: agents building mobile apps | Not measured | 1090 | Only GPT-5.2-Codex measured |
| Design Arena: agents building Android apps | Not measured | 1176 | Only GPT-5.2-Codex measured |
| Design Arena: agents building Godot games | Not measured | 1142 | Only GPT-5.2-Codex measured |
| τ²-Bench Telecom | 47.1% | 92.1% | GPT-5.2-Codex |
| Neither model has a result on the other 33 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | GPT-4.1 | GPT-5.2-Codex | Leader |
|---|---|---|---|
| GPQA Diamond | 65.4% | Not measured | Only GPT-4.1 measured |
| MMMU Pro (Vals AI) | 72.4% | Not measured | Only GPT-4.1 measured |
| GPQA Diamond (Artificial Analysis run) | 66.6% | 89.9% | GPT-5.2-Codex |
| Humanity's Last Exam (Artificial Analysis run) | 4.2% | 35.7% | GPT-5.2-Codex |
| GPQA Diamond (Epoch AI run) | 66.9% | Not measured | Only GPT-4.1 measured |
| ARC-AGI | 5.5% | Not measured | Only GPT-4.1 measured |
| ARC-AGI-2 | 0.4% | Not measured | Only GPT-4.1 measured |
| SimpleBench | 27.0% | Not measured | Only GPT-4.1 measured |
| Chess puzzles | 3.0% | Not measured | Only GPT-4.1 measured |
| LMArena Text: hard prompts | 1384 | Not measured | Only GPT-4.1 measured |
| LMArena Text: expert | 1364 | Not measured | Only GPT-4.1 measured |
| LMArena Text: hard prompts in English | 1393 | Not measured | Only GPT-4.1 measured |
| LMArena Vision: diagrams | 1215 | Not measured | Only GPT-4.1 measured |
| LMArena Vision: homework | 1248 | Not measured | Only GPT-4.1 measured |
| Neither model has a result on the other 7 boards here | |||
MathDirectional onlyPublic boards, shown as published, kept out of the index10 boards, none shared
| Benchmark | GPT-4.1 | GPT-5.2-Codex | Leader |
|---|---|---|---|
| AIME | 39.6% | Not measured | Only GPT-4.1 measured |
| MATH 500 (Vals AI) | 87.2% | Not measured | Only GPT-4.1 measured |
| MGSM (Vals AI) | 87.7% | Not measured | Only GPT-4.1 measured |
| AA Math Index | 34.7 | Not measured | Only GPT-4.1 measured |
| AIME 2025 (Artificial Analysis run) | 34.7% | Not measured | Only GPT-4.1 measured |
| OTIS Mock AIME | 38.3% | Not measured | Only GPT-4.1 measured |
| FrontierMath, tiers 1 to 3 | 6.0% | Not measured | Only GPT-4.1 measured |
| MATH level 5 | 83.0% | Not measured | Only GPT-4.1 measured |
| LMArena Text: math | 1369 | Not measured | Only GPT-4.1 measured |
| LMArena Text: mathematical work | 1360 | Not measured | Only GPT-4.1 measured |
| Neither model has a result on the other 3 boards here | |||
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index14 boards, none shared
| Benchmark | GPT-4.1 | GPT-5.2-Codex | Leader |
|---|---|---|---|
| MMLU Pro | 80.5% | Not measured | Only GPT-4.1 measured |
| LegalBench (Vals AI) | 83.1% | Not measured | Only GPT-4.1 measured |
| CaseLaw (Vals AI) | 69.9% | Not measured | Only GPT-4.1 measured |
| CorpFin (Vals AI) | 63.0% | Not measured | Only GPT-4.1 measured |
| TaxEval (Vals AI) | 75.1% | Not measured | Only GPT-4.1 measured |
| MortgageTax (Vals AI) | 65.9% | Not measured | Only GPT-4.1 measured |
| MedQA (Vals AI) | 91.2% | Not measured | Only GPT-4.1 measured |
| MMLU Pro (Artificial Analysis run) | 80.6% | Not measured | Only GPT-4.1 measured |
| SimpleQA Verified | 31.1% | Not measured | Only GPT-4.1 measured |
| GeoBench | 72.0% | Not measured | Only GPT-4.1 measured |
| LMArena Text: science | 1382 | Not measured | Only GPT-4.1 measured |
| LMArena Text: medicine and healthcare | 1379 | Not measured | Only GPT-4.1 measured |
| LMArena Text: legal and government | 1399 | Not measured | Only GPT-4.1 measured |
| LMArena Text: business and finance | 1369 | Not measured | Only GPT-4.1 measured |
| Neither model has a result on the other 3 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index6 boards, none shared
| Benchmark | GPT-4.1 | GPT-5.2-Codex | Leader |
|---|---|---|---|
| Design Arena: websites | 1039 | Not measured | Only GPT-4.1 measured |
| Design Arena: UI components | 1000 | Not measured | Only GPT-4.1 measured |
| Design Arena: game dev | 1080 | Not measured | Only GPT-4.1 measured |
| Design Arena: data viz | 1104 | Not measured | Only GPT-4.1 measured |
| Design Arena: 3D | 850 | Not measured | Only GPT-4.1 measured |
| Design Arena: all code categories | 1026 | Not measured | Only GPT-4.1 measured |
| Neither model has a result on the other 3 boards here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | GPT-4.1 | GPT-5.2-Codex | Leader |
|---|---|---|---|
| IFBench | 43.0% | 77.6% | GPT-5.2-Codex |
| Long context reasoning (AA-LCR) | 68.3% | 82.3% | GPT-5.2-Codex |
| DTBench | 68.3% | Not measured | Only GPT-4.1 measured |
| LMCA | 25.6% | Not measured | Only GPT-4.1 measured |
| Fiction.LiveBench | 63.9% | Not measured | Only GPT-4.1 measured |
| LMArena Text | 1383 | Not measured | Only GPT-4.1 measured |
| LMArena Text: instruction following | 1367 | Not measured | Only GPT-4.1 measured |
| LMArena Text: creative writing | 1363 | Not measured | Only GPT-4.1 measured |
| LMArena Vision | 1210 | Not measured | Only GPT-4.1 measured |
| LMArena Text: English | 1392 | Not measured | Only GPT-4.1 measured |
| LMArena Text: non-English | 1370 | Not measured | Only GPT-4.1 measured |
| LMArena Text: Chinese | 1382 | Not measured | Only GPT-4.1 measured |
| LMArena Text: German | 1381 | Not measured | Only GPT-4.1 measured |
| LMArena Text: Japanese | 1319 | Not measured | Only GPT-4.1 measured |
| LMArena Text: Korean | 1338 | Not measured | Only GPT-4.1 measured |
| LMArena Text: French | 1382 | Not measured | Only GPT-4.1 measured |
| LMArena Text: Spanish | 1376 | Not measured | Only GPT-4.1 measured |
| LMArena Text: Russian | 1377 | Not measured | Only GPT-4.1 measured |
| LMArena Text: Polish | 1389 | Not measured | Only GPT-4.1 measured |
| LMArena Text: multi-turn | 1397 | Not measured | Only GPT-4.1 measured |
| LMArena Text: longer queries | 1385 | Not measured | Only GPT-4.1 measured |
| LMArena Text: writing and language | 1371 | Not measured | Only GPT-4.1 measured |
| LMArena Text: entertainment and media | 1365 | Not measured | Only GPT-4.1 measured |
| LMArena Vision: English | 1209 | Not measured | Only GPT-4.1 measured |
| LMArena Vision: Chinese | 1226 | Not measured | Only GPT-4.1 measured |
| LMArena Vision: reading text | 1218 | Not measured | Only GPT-4.1 measured |
| LMArena Vision: humor | 1196 | Not measured | Only GPT-4.1 measured |
| LMArena Vision: captioning | 1207 | Not measured | Only GPT-4.1 measured |
| LMArena Vision: recognizing things | 1208 | Not measured | Only GPT-4.1 measured |
| LMArena Vision: creative writing | 1210 | Not measured | Only GPT-4.1 measured |
| Neither model has a result on the other 16 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | GPT-4.1 | GPT-5.2-Codex | 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 costs less
Repository review
50K input + 3K output tokens
GPT-4.1 costs less
Agent loop
200K input + 10K output tokens
GPT-4.1 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.
| Specification | OpenAI GPT-4.1 | OpenAI GPT-5.2-Codex |
|---|---|---|
| Maker | OpenAI | OpenAI |
| Weights | No public weights listed | No public weights listed |
| Released | April 14, 2025 | January 14, 2026 |
| Context window | 1.05M | 400K |
| Input $/Mtok | $2.00 | $1.75 |
| Output $/Mtok | $8.00 | $14.00 |
| Coding index | 46.5 | 55.1 |
| Independent measurements behind it | 2 | 2 |
| Cost per solved task | $0.37 | $0.43 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is GPT-4.1 better than GPT-5.2-Codex for coding?
Pick GPT-5.2-Codex for coding: it scores 55.1 on the Vibe Coding Index against 46.5 for GPT-4.1. Pick GPT-4.1 when output tokens cost 1.8x less, or when a solved task costs $0.37 against $0.43.
Which is cheaper, GPT-4.1 or GPT-5.2-Codex?
GPT-4.1 costs $2.00 per million input tokens and $8.00 per million output tokens; GPT-5.2-Codex costs $1.75 and $14.00. Per solved SWE-bench Verified task, GPT-4.1 costs $0.37 and GPT-5.2-Codex $0.43.
Which has the bigger context window, GPT-4.1 or GPT-5.2-Codex?
GPT-4.1, with 1.05M tokens against 400K.
How many benchmarks is the GPT-4.1 vs GPT-5.2-Codex comparison based on?
10 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-5.2-Codex's on 2. Where only one model has a result, the page says so instead of naming a winner.