# Claude Opus 4.5 vs GPT-5

> Claude Opus 4.5 and GPT-5 score almost the same on the Noometry Index (50.5 vs 50.9), so choose on price, context window or the category you care about most.

- Canonical page: https://noometry.com/compare/claude-opus-4-5-vs-gpt-5
- Last updated: 2026-10-10
- Shared benchmarks: 56

## Summary

- They share 56 benchmarks with published results for both. Claude Opus 4.5 scores higher in 6 categories and GPT-5 in 4 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in long context, where GPT-5 leads 69.5 to 46.4.
- The biggest single-benchmark swing is ARC-AGI-2: 37.6% for Claude Opus 4.5 and 9.9% for GPT-5.
- GPT-5 is cheaper at $1.25 / $10 per million input/output tokens, against $5 / $25 for Claude Opus 4.5.
- GPT-5 accepts more context: 400K tokens versus 200K.

## Snapshot

| | Claude Opus 4.5 | GPT-5 |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 50.5 | 50.9 |
| Rank | 39 | 37 |
| Context | 200K | 400K |
| Input $/M | $5 | $1.25 |
| Output $/M | $25 | $10 |
| Weights | Proprietary | Proprietary |

## Coding

- Claude Opus 4.5: 55.4 (#21)
- GPT-5: 50.7 (#39)

| Benchmark | Claude Opus 4.5 | GPT-5 |
|---|---|---|
| SWE-bench Verified | 76.7% | 73.6% |
| SWE-bench Verified (bash only) | 76.8% | 65% |
| LMArena WebDev | 1493 | 1417 |
| GSO | 26.5% | 6.9% |
| WeirdML | 63.7% | 60.7% |
| LMArena Coding | 1504 | 1436 |
| ALE-Bench | 1,025 | 1,162 |
| AlgoTune | 1.77 | 1.67 |
| Aider Polyglot | — | 88% |
| SWE-bench Multilingual | 70.7% | — |
| SciCode | — | 42.9% |

## Agentic & Tool Use

- Claude Opus 4.5: 47.2 (#11)
- GPT-5: 33.0 (#53)

| Benchmark | Claude Opus 4.5 | GPT-5 |
|---|---|---|
| Terminal-Bench | 63.1% | 49.6% |
| GDPval | 45.5% | 34.8% |
| Remote Labor Index | 3.8% | 1.7% |
| DeepResearch Bench | 54.8% | 49.6% |
| BALROG | 43.5% | 32.8% |
| LMArena Search | 1180 | 1133 |
| METR Time Horizons | 75% | 69.6% |
| Berkeley Function Calling Leaderboard | 77.5% | — |
| τ²-bench Airline | 84% | — |
| τ²-bench Banking | 24.7% | — |
| τ²-bench Retail | 79.6% | — |
| τ²-bench Telecom | 92.3% | — |
| Cybench | 82% | — |
| OSWorld | 66.3% | — |
| Vending-Bench 2 | 4,967 | — |

## Reasoning

- Claude Opus 4.5: 42.5 (#43)
- GPT-5: 38.3 (#55)

| Benchmark | Claude Opus 4.5 | GPT-5 |
|---|---|---|
| ARC-AGI-2 | 37.6% | 9.9% |
| SimpleBench | 62% | 56.7% |
| Kagi LLM Benchmark | 80.2% | 72.7% |
| ARC-AGI-1 | 80% | 65.7% |
| Chess Puzzles | 12% | 37% |
| EnigmaEval | 11.9% | 10.5% |
| EBR-Bench | 14.3% | 12.7% |
| LMArena Hard Prompts | 1476 | 1415 |
| Mystery Game Puzzles | 22% | 23% |
| DTBench | 89.9% | 90.7% |
| LMCA | 44.5% | 40% |
| Epoch Capabilities Index | 150.1 | 150 |
| ForecastBench | 60.7 | 61.4 |
| NYT Connections (extended) | 52.5% | — |
| CritPt | — | 12.6% |

## Math

- Claude Opus 4.5: 38.5 (#85)
- GPT-5: 54.6 (#36)

| Benchmark | Claude Opus 4.5 | GPT-5 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 34.4% | 55.4% |
| FrontierMath Tier 4 | 4.9% | 22% |
| OTIS Mock AIME 2024-2025 | 86.1% | 91.4% |
| ProofBench | 36% | 18% |
| LMArena Math | 1463 | 1406 |
| FrontierMath (Feb 2025 set) | 20.7% | 32.4% |
| FrontierMath Tier 4 (v1) | 4.2% | 12.5% |
| Omni-MATH | — | 64.7% |
| MATH Level 5 | — | 98.1% |

## Knowledge

- Claude Opus 4.5: 56.4 (#38)
- GPT-5: 56.5 (#37)

| Benchmark | Claude Opus 4.5 | GPT-5 |
|---|---|---|
| GPQA Diamond | 86% | 86.2% |
| Humanity's Last Exam | 25.2% | 25.3% |
| SimpleQA Verified | 45.7% | 50.1% |
| Vectara Hallucination Rate | 10.9% | 14.7% |
| LMArena Expert | 1488 | 1420 |
| MMLU-Pro | — | 86.3% |
| Confabulations | — | 10.3% |
| GPQA (HELM) | — | 79.2% |

## Multimodal

- Claude Opus 4.5: 31.1 (#99)
- GPT-5: 46.1 (#12)

| Benchmark | Claude Opus 4.5 | GPT-5 |
|---|---|---|
| GeoBench | 75% | 81% |
| VPCT | 40% | 66% |
| LMArena Vision | — | 1232 |
| Furniture Assembly | 28.3% | — |
| LMArena Document | 1462 | — |

## Multilingual

- Claude Opus 4.5: 54.5 (#45)
- GPT-5: 51.5 (#105)

| Benchmark | Claude Opus 4.5 | GPT-5 |
|---|---|---|
| LMArena Non-English | 1438 | 1397 |
| LMArena Chinese | 1470 | 1423 |
| LMArena French | 1471 | 1409 |
| LMArena German | 1449 | 1415 |
| LMArena Japanese | 1415 | 1411 |
| LMArena Korean | 1424 | 1361 |
| LMArena Russian | 1448 | 1406 |
| LMArena Spanish | 1455 | 1398 |

## Instruction Following

- Claude Opus 4.5: 77.7 (#20)
- GPT-5: 74.0 (#107)

| Benchmark | Claude Opus 4.5 | GPT-5 |
|---|---|---|
| LMArena Instruction Following | 1478 | 1388 |
| IFEval | — | 87.5% |

## Long Context

- Claude Opus 4.5: 46.4 (#21)
- GPT-5: 69.5 (#2)

| Benchmark | Claude Opus 4.5 | GPT-5 |
|---|---|---|
| LMArena Longer Query | 1480 | 1399 |
| Fiction.LiveBench | — | 97.2% |
| CL-bench | 21.1% | — |

## Writing & Preference

- Claude Opus 4.5: 68.2 (#28)
- GPT-5: 63.5 (#63)

| Benchmark | Claude Opus 4.5 | GPT-5 |
|---|---|---|
| LMArena Text | 1451 | 1406 |
| LMArena Creative Writing | 1445 | 1365 |
| EQ-Bench Creative Writing | 1687 | 1627 |
| LMArena Multi-Turn | 1466 | 1426 |
| Short-Story Creative Writing | — | 86% |
| WildBench | — | 85.7% |

## FAQ

### Is Claude Opus 4.5 better than GPT-5?

Claude Opus 4.5 and GPT-5 score almost the same on the Noometry Index (50.5 vs 50.9), so choose on price, context window or the category you care about most.

### Which is cheaper, Claude Opus 4.5 or GPT-5?

GPT-5 is cheaper. It lists at $1.25 per million input tokens and $10 per million output tokens; Claude Opus 4.5 lists at $5 and $25.

### Is Claude Opus 4.5 or GPT-5 better for coding?

Claude Opus 4.5 scores higher on coding benchmarks: 55.4 versus 50.7 in the Noometry coding category.

### Which has the bigger context window?

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

### How many benchmarks do Claude Opus 4.5 and GPT-5 share?

56 benchmarks have published results for both models. Claude Opus 4.5 has 69 scored results on Noometry and GPT-5 has 69.