# Claude 3.7 Sonnet vs GPT-5.4

> GPT-5.4 is the stronger model overall, scoring 59.5 to 39.4 on the Noometry Index.

- Canonical page: https://noometry.com/compare/claude-3-7-sonnet-vs-gpt-5-4
- Last updated: 2026-10-10
- Shared benchmarks: 32

## Summary

- They share 32 benchmarks with published results for both. Claude 3.7 Sonnet scores higher in 0 categories and GPT-5.4 in 10 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.4 leads 61.7 to 18.7.
- The biggest single-benchmark swing is ARC-AGI-2: 0.9% for Claude 3.7 Sonnet and 74% for GPT-5.4.

## Snapshot

| | Claude 3.7 Sonnet | GPT-5.4 |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 39.4 | 59.5 |
| Rank | 154 | 14 |
| Context | — | 1.05M |
| Input $/M | — | $2.50 |
| Output $/M | — | $15 |
| Weights | Proprietary | Proprietary |

## Coding

- Claude 3.7 Sonnet: 41.1 (#120)
- GPT-5.4: 52.8 (#28)

| Benchmark | Claude 3.7 Sonnet | GPT-5.4 |
|---|---|---|
| SWE-bench Verified | 61% | 76.9% |
| GSO | 3.8% | 31.4% |
| LMArena Coding | 1361 | 1497 |
| DeepSWE | — | 51.8% |
| SWE-bench Verified (bash only) | 52.8% | — |
| Aider Polyglot | 64.9% | — |
| LMArena WebDev | — | 1465 |
| SciCode | — | 56.6% |
| WeirdML | — | 77.7% |
| LiveBench Coding | 74.5% | — |
| MirrorCode | — | 15.6% |
| CadEval | 54% | — |
| ALE-Bench | — | 1,607 |
| AlgoTune | — | 1.85 |

## Agentic & Tool Use

- Claude 3.7 Sonnet: 33.9 (#46)
- GPT-5.4: 46.5 (#12)

| Benchmark | Claude 3.7 Sonnet | GPT-5.4 |
|---|---|---|
| DeepResearch Bench | 43.6% | 35.1% |
| METR Time Horizons | 60% | 74.3% |
| Terminal-Bench | — | 81.8% |
| APEX-Agents | — | 52.4% |
| TheAgentCompany | 30.9% | — |
| τ²-bench Banking | — | 39.4% |
| Cybench | 20% | — |
| OSWorld | 35.8% | — |
| PostTrainBench | — | 19% |
| GBAEval | — | 45.1% |
| LMArena Search | — | 1197 |
| Vending-Bench 2 | — | 6,144 |

## Reasoning

- Claude 3.7 Sonnet: 18.7 (#252)
- GPT-5.4: 61.7 (#17)

| Benchmark | Claude 3.7 Sonnet | GPT-5.4 |
|---|---|---|
| ARC-AGI-2 | 0.9% | 74% |
| ARC-AGI-1 | 28.6% | 93.7% |
| EnigmaEval | 4.2% | 16% |
| LMArena Hard Prompts | 1333 | 1485 |
| Epoch Capabilities Index | 141.16 | 156.92 |
| ForecastBench | 61.8 | 59.5 |
| SimpleBench | 46.4% | — |
| Kagi LLM Benchmark | — | 63.8% |
| NYT Connections (extended) | — | 91.3% |
| CritPt | — | 23.4% |
| Chess Puzzles | — | 44% |
| Thematic Generalization | — | 80% |
| EBR-Bench | — | 25.4% |
| LiveBench Reasoning | 87.8% | — |
| Mystery Game Puzzles | — | 37% |
| DTBench | — | 94.4% |
| LiveBench Data Analysis | 74% | — |
| LMCA | — | 52% |
| LiveBench | 76.1% | — |

## Math

- Claude 3.7 Sonnet: 35.7 (#137)
- GPT-5.4: 73.3 (#14)

| Benchmark | Claude 3.7 Sonnet | GPT-5.4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 57.8% | 97.8% |
| LMArena Math | 1336 | 1488 |
| FrontierMath (Feb 2025 set) | 4.1% | 47.6% |
| FrontierMath (Tiers 1-3) | — | 78.6% |
| FrontierMath Tier 4 | — | 49% |
| MathArena Final-Answer Competitions | — | 83.1% |
| ProofBench | — | 56% |
| Omni-MATH | 33% | — |
| LiveBench Math | 79% | — |
| MATH Level 5 | 91.2% | — |
| FrontierMath Tier 4 (v1) | — | 27.1% |

## Knowledge

- Claude 3.7 Sonnet: 39.7 (#117)
- GPT-5.4: 65.3 (#13)

| Benchmark | Claude 3.7 Sonnet | GPT-5.4 |
|---|---|---|
| GPQA Diamond | 79.7% | 93.3% |
| Humanity's Last Exam | 8% | 36.2% |
| LMArena Expert | 1321 | 1507 |
| SimpleQA Verified | — | 45.1% |
| MMLU-Pro | 78.4% | — |
| Confabulations | 14.7% | — |
| Vectara Hallucination Rate | — | 7% |
| GPQA (HELM) | 60.8% | — |

## Multimodal

- Claude 3.7 Sonnet: 33.1 (#90)
- GPT-5.4: 43.8 (#19)

| Benchmark | Claude 3.7 Sonnet | GPT-5.4 |
|---|---|---|
| LMArena Vision | 1169 | 1302 |
| GeoBench | 68% | — |
| VPCT | 39% | — |
| Blueprint-Bench 2 | — | 27.1% |
| Furniture Assembly | — | 37.5% |
| LMArena Document | — | 1471 |
| SpatialViz-Bench | 33.9% | — |

## Multilingual

- Claude 3.7 Sonnet: 44.3 (#174)
- GPT-5.4: 56.4 (#22)

| Benchmark | Claude 3.7 Sonnet | GPT-5.4 |
|---|---|---|
| LMArena Non-English | 1296 | 1465 |
| LMArena Chinese | 1299 | 1520 |
| LMArena French | 1302 | 1493 |
| LMArena German | 1301 | 1472 |
| LMArena Japanese | 1268 | 1484 |
| LMArena Korean | 1249 | 1449 |
| LMArena Russian | 1311 | 1480 |
| LMArena Spanish | 1299 | 1453 |

## Instruction Following

- Claude 3.7 Sonnet: 73.1 (#120)
- GPT-5.4: 77.3 (#25)

| Benchmark | Claude 3.7 Sonnet | GPT-5.4 |
|---|---|---|
| LMArena Instruction Following | 1352 | 1469 |
| LiveBench Instruction Following | 81.3% | — |
| IFEval | 83.4% | — |

## Long Context

- Claude 3.7 Sonnet: 50.2 (#10)
- GPT-5.4: 50.3 (#8)

| Benchmark | Claude 3.7 Sonnet | GPT-5.4 |
|---|---|---|
| LMArena Longer Query | 1373 | 1474 |
| Fiction.LiveBench | 83.3% | — |
| CL-bench | — | 27.9% |
| CL-bench Life | — | 21.7% |

## Writing & Preference

- Claude 3.7 Sonnet: 54.6 (#145)
- GPT-5.4: 72.1 (#18)

| Benchmark | Claude 3.7 Sonnet | GPT-5.4 |
|---|---|---|
| LMArena Text | 1314 | 1469 |
| LMArena Creative Writing | 1332 | 1439 |
| EQ-Bench Creative Writing | 1412 | 1840 |
| LMArena Multi-Turn | 1339 | 1482 |
| Short-Story Creative Writing | 81.1% | — |
| WildBench | 81.4% | — |
| EQ-Bench 4 | — | 1272 |
| LiveBench Language | 59.9% | — |

## FAQ

### Is Claude 3.7 Sonnet better than GPT-5.4?

GPT-5.4 is the stronger model overall, scoring 59.5 to 39.4 on the Noometry Index.

### Is Claude 3.7 Sonnet or GPT-5.4 better for coding?

GPT-5.4 scores higher on coding benchmarks: 52.8 versus 41.1 in the Noometry coding category.

### How many benchmarks do Claude 3.7 Sonnet and GPT-5.4 share?

32 benchmarks have published results for both models. Claude 3.7 Sonnet has 58 scored results on Noometry and GPT-5.4 has 68.