# GPT-4.1 vs GPT-4.5

> GPT-4.5 is the stronger model overall, scoring 36.9 to 35.7 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gpt-4-1-vs-gpt-4-5
- Last updated: 2026-10-10
- Shared benchmarks: 31

## Summary

- They share 31 benchmarks with published results for both. GPT-4.1 scores higher in 4 categories and GPT-4.5 in 6 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-4.5 leads 31.0 to 21.4.
- The biggest single-benchmark swing is Aider Polyglot: 52.4% for GPT-4.1 and 44.9% for GPT-4.5.

## Snapshot

| | GPT-4.1 | GPT-4.5 |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 35.7 | 36.9 |
| Rank | 215 | 200 |
| Context | 1.05M | — |
| Input $/M | $2 | — |
| Output $/M | $8 | — |
| Weights | Proprietary | Proprietary |

## Coding

- GPT-4.1: 34.2 (#228)
- GPT-4.5: 41.8 (#102)

| Benchmark | GPT-4.1 | GPT-4.5 |
|---|---|---|
| Aider Polyglot | 52.4% | 44.9% |
| WeirdML | 39% | 39.4% |
| LMArena Coding | 1391 | 1397 |
| SWE-bench Verified | 48.5% | — |
| SWE-bench Verified (bash only) | 39.6% | — |
| LiveBench Coding | — | 75.2% |
| CadEval | 42% | — |
| ALE-Bench | 558.1 | — |

## Agentic & Tool Use

- GPT-4.1: 34.4 (#40)
- GPT-4.5: 27.6 (#92)

| Benchmark | GPT-4.1 | GPT-4.5 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 54% | — |
| Cybench | — | 17.5% |

## Reasoning

- GPT-4.1: 11.7 (#334)
- GPT-4.5: 14.1 (#320)

| Benchmark | GPT-4.1 | GPT-4.5 |
|---|---|---|
| ARC-AGI-2 | 0.4% | 0.8% |
| SimpleBench | 27% | 34.5% |
| ARC-AGI-1 | 5.5% | 10.3% |
| EnigmaEval | 2.2% | 3.2% |
| LMArena Hard Prompts | 1384 | 1404 |
| Epoch Capabilities Index | 136.79 | 136.74 |
| ForecastBench | 61.5 | 61.7 |
| Kagi LLM Benchmark | 52.3% | — |
| Chess Puzzles | 6% | — |
| LiveBench Reasoning | — | 71.1% |
| DTBench | 68.3% | — |
| LiveBench Data Analysis | — | 64.3% |
| LMCA | 25.6% | — |
| LiveBench | — | 69% |

## Math

- GPT-4.1: 21.4 (#269)
- GPT-4.5: 31.0 (#200)

| Benchmark | GPT-4.1 | GPT-4.5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 38.3% | 37.8% |
| LMArena Math | 1369 | 1412 |
| MATH Level 5 | 83% | 78.6% |
| FrontierMath (Tiers 1-3) | 6% | — |
| Omni-MATH | 47.1% | — |
| LiveBench Math | — | 69.3% |
| FrontierMath (Feb 2025 set) | 5.5% | — |
| FrontierMath Tier 4 (v1) | 0% | — |

## Knowledge

- GPT-4.1: 37.0 (#152)
- GPT-4.5: 32.4 (#205)

| Benchmark | GPT-4.1 | GPT-4.5 |
|---|---|---|
| GPQA Diamond | 66.9% | 68.7% |
| Humanity's Last Exam | 5.4% | 5.4% |
| LMArena Expert | 1364 | 1394 |
| SimpleQA Verified | 31.1% | — |
| MMLU-Pro | 81.1% | — |
| Confabulations | — | 13.6% |
| Vectara Hallucination Rate | 5.6% | — |
| GPQA (HELM) | 65.9% | — |

## Multimodal

- GPT-4.1: 37.5 (#66)
- GPT-4.5: 37.0 (#68)

| Benchmark | GPT-4.1 | GPT-4.5 |
|---|---|---|
| LMArena Vision | 1210 | 1195 |
| GeoBench | 72% | — |
| VPCT | — | 45% |

## Multilingual

- GPT-4.1: 49.6 (#128)
- GPT-4.5: 52.7 (#78)

| Benchmark | GPT-4.1 | GPT-4.5 |
|---|---|---|
| LMArena Non-English | 1370 | 1413 |
| LMArena Chinese | 1382 | 1421 |
| LMArena French | 1383 | 1417 |
| LMArena German | 1380 | 1457 |
| LMArena Japanese | 1320 | 1416 |
| LMArena Korean | 1339 | 1392 |
| LMArena Russian | 1377 | 1419 |
| LMArena Spanish | 1375 | — |

## Instruction Following

- GPT-4.1: 71.5 (#148)
- GPT-4.5: 72.8 (#129)

| Benchmark | GPT-4.1 | GPT-4.5 |
|---|---|---|
| LMArena Instruction Following | 1367 | 1404 |
| LiveBench Instruction Following | — | 72.3% |
| IFEval | 83.8% | — |

## Long Context

- GPT-4.1: 40.0 (#158)
- GPT-4.5: 40.4 (#150)

| Benchmark | GPT-4.1 | GPT-4.5 |
|---|---|---|
| Fiction.LiveBench | 63.9% | 63.9% |
| LMArena Longer Query | 1385 | 1406 |

## Writing & Preference

- GPT-4.1: 57.8 (#120)
- GPT-4.5: 57.1 (#129)

| Benchmark | GPT-4.1 | GPT-4.5 |
|---|---|---|
| LMArena Text | 1383 | 1417 |
| LMArena Creative Writing | 1363 | 1394 |
| EQ-Bench Creative Writing | 1420 | 1258 |
| LMArena Multi-Turn | 1398 | 1444 |
| Short-Story Creative Writing | — | 75.6% |
| WildBench | 85.4% | — |
| LiveBench Language | — | 61.5% |

## FAQ

### Is GPT-4.1 better than GPT-4.5?

GPT-4.5 is the stronger model overall, scoring 36.9 to 35.7 on the Noometry Index.

### Is GPT-4.1 or GPT-4.5 better for coding?

GPT-4.5 scores higher on coding benchmarks: 41.8 versus 34.2 in the Noometry coding category.

### How many benchmarks do GPT-4.1 and GPT-4.5 share?

31 benchmarks have published results for both models. GPT-4.1 has 52 scored results on Noometry and GPT-4.5 has 42.