# Claude 3.5 Sonnet vs GPT-4

> Claude 3.5 Sonnet is the stronger model overall, scoring 34.4 to 28.9 on the Noometry Index.

- Canonical page: https://noometry.com/compare/claude-3-5-sonnet-vs-gpt-4
- Last updated: 2026-10-10
- Shared benchmarks: 30

## Summary

- They share 30 benchmarks with published results for both. Claude 3.5 Sonnet scores higher in 8 categories and GPT-4 in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Claude 3.5 Sonnet leads 53.0 to 35.0.
- The biggest single-benchmark swing is MATH Level 5: 56.9% for Claude 3.5 Sonnet and 23% for GPT-4.

## Snapshot

| | Claude 3.5 Sonnet | GPT-4 |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 34.4 | 28.9 |
| Rank | 225 | 308 |
| Context | — | 8K |
| Input $/M | — | $30 |
| Output $/M | — | $60 |
| Weights | Proprietary | Proprietary |

## Coding

- Claude 3.5 Sonnet: 39.5 (#145)
- GPT-4: 31.2 (#277)

| Benchmark | Claude 3.5 Sonnet | GPT-4 |
|---|---|---|
| WeirdML | 40% | 12.4% |
| BigCodeBench Instruct | 46.8% | 46% |
| LMArena Coding | 1343 | 1254 |
| BigCodeBench Complete | 58.6% | 57.2% |
| HumanEval+ | 81.7% | 79.3% |
| Aider Polyglot | 51.6% | — |
| GSO | 4.6% | — |
| LiveBench Coding | 67.1% | — |
| CadEval | 48% | — |
| MBPP+ | 74.3% | — |

## Agentic & Tool Use

- Claude 3.5 Sonnet: 32.1 (#61)
- GPT-4: —

| Benchmark | Claude 3.5 Sonnet | GPT-4 |
|---|---|---|
| METR Time Horizons | 45.2% | 36.1% |
| TheAgentCompany | 24% | — |
| Cybench | 17.5% | — |
| BALROG | 32.6% | — |

## Reasoning

- Claude 3.5 Sonnet: 22.5 (#176)
- GPT-4: 17.3 (#285)

| Benchmark | Claude 3.5 Sonnet | GPT-4 |
|---|---|---|
| LMArena Hard Prompts | 1305 | 1241 |
| DTBench | 67.8% | 62.7% |
| Epoch Capabilities Index | 133.55 | 125.88 |
| ForecastBench | 60.7 | 57.8 |
| SimpleBench | 41.4% | — |
| Chess Puzzles | — | 4% |
| EnigmaEval | 0.9% | — |
| LiveBench Reasoning | 56.7% | — |
| Mystery Game Puzzles | — | 12% |
| LiveBench Data Analysis | 55% | — |
| LMCA | — | 17.1% |
| BIG-Bench Hard | — | 75.1% |
| HellaSwag | — | 95.3% |
| LiveBench | 59% | — |
| WinoGrande | — | 87.5% |

## Math

- Claude 3.5 Sonnet: 18.0 (#282)
- GPT-4: 10.1 (#304)

| Benchmark | Claude 3.5 Sonnet | GPT-4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 8.5% | 1.1% |
| LMArena Math | 1306 | 1269 |
| MATH Level 5 | 56.9% | 23% |
| Omni-MATH | 27.6% | — |
| LiveBench Math | 52.3% | — |
| FrontierMath (Feb 2025 set) | 2.1% | — |
| FrontierMath Tier 4 (v1) | 0% | — |
| GSM8K | — | 92% |

## Knowledge

- Claude 3.5 Sonnet: 28.5 (#236)
- GPT-4: 18.4 (#277)

| Benchmark | Claude 3.5 Sonnet | GPT-4 |
|---|---|---|
| GPQA Diamond | 55.3% | 35.7% |
| LMArena Expert | 1264 | 1211 |
| MMLU | 87.3% | 86.4% |
| Humanity's Last Exam | 4.1% | — |
| MMLU-Pro | 77.7% | — |
| Confabulations | 19.9% | — |
| GPQA (HELM) | 56.5% | — |
| TriviaQA | — | 84.8% |

## Multimodal

- Claude 3.5 Sonnet: 26.0 (#115)
- GPT-4: —

| Benchmark | Claude 3.5 Sonnet | GPT-4 |
|---|---|---|
| LMArena Vision | 1125 | — |
| Video-MME | 60% | — |
| GeoBench | 62% | — |
| VPCT | 33% | — |

## Multilingual

- Claude 3.5 Sonnet: 43.4 (#180)
- GPT-4: 40.8 (#210)

| Benchmark | Claude 3.5 Sonnet | GPT-4 |
|---|---|---|
| LMArena Non-English | 1283 | 1246 |
| LMArena Chinese | 1271 | 1242 |
| LMArena French | 1304 | 1282 |
| LMArena German | 1297 | 1251 |
| LMArena Japanese | 1234 | 1209 |
| LMArena Korean | 1200 | 1184 |
| LMArena Russian | 1305 | 1251 |
| LMArena Spanish | 1289 | 1261 |

## Instruction Following

- Claude 3.5 Sonnet: 69.0 (#177)
- GPT-4: 65.5 (#217)

| Benchmark | Claude 3.5 Sonnet | GPT-4 |
|---|---|---|
| LMArena Instruction Following | 1297 | 1241 |
| LiveBench Instruction Following | 69.3% | — |
| IFEval | 85.5% | — |

## Long Context

- Claude 3.5 Sonnet: 39.9 (#162)
- GPT-4: 37.7 (#207)

| Benchmark | Claude 3.5 Sonnet | GPT-4 |
|---|---|---|
| LMArena Longer Query | 1311 | 1244 |

## Writing & Preference

- Claude 3.5 Sonnet: 53.0 (#159)
- GPT-4: 35.0 (#263)

| Benchmark | Claude 3.5 Sonnet | GPT-4 |
|---|---|---|
| LMArena Text | 1298 | 1263 |
| LMArena Creative Writing | 1292 | 1244 |
| EQ-Bench Creative Writing | 1451 | 752 |
| LMArena Multi-Turn | 1326 | 1257 |
| Short-Story Creative Writing | 80.3% | — |
| WildBench | 79.2% | — |
| LiveBench Language | 53.8% | — |

## FAQ

### Is Claude 3.5 Sonnet better than GPT-4?

Claude 3.5 Sonnet is the stronger model overall, scoring 34.4 to 28.9 on the Noometry Index.

### Is Claude 3.5 Sonnet or GPT-4 better for coding?

Claude 3.5 Sonnet scores higher on coding benchmarks: 39.5 versus 31.2 in the Noometry coding category.

### How many benchmarks do Claude 3.5 Sonnet and GPT-4 share?

30 benchmarks have published results for both models. Claude 3.5 Sonnet has 60 scored results on Noometry and GPT-4 has 38.