# Claude 3 Haiku vs GPT-4

> GPT-4 is the stronger model overall, scoring 28.9 to 25.7 on the Noometry Index.

- Canonical page: https://noometry.com/compare/claude-3-haiku-vs-gpt-4
- Last updated: 2026-10-10
- Shared benchmarks: 31

## Summary

- They share 31 benchmarks with published results for both. Claude 3 Haiku scores higher in 0 categories and GPT-4 in 8 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where GPT-4 leads 35.0 to 29.8.
- The biggest single-benchmark swing is DTBench: 50.1% for Claude 3 Haiku and 62.7% for GPT-4.

## Snapshot

| | Claude 3 Haiku | GPT-4 |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 25.7 | 28.9 |
| Rank | 333 | 308 |
| Context | — | 8K |
| Input $/M | — | $30 |
| Output $/M | — | $60 |
| Weights | Proprietary | Proprietary |

## Coding

- Claude 3 Haiku: 26.1 (#320)
- GPT-4: 31.2 (#277)

| Benchmark | Claude 3 Haiku | GPT-4 |
|---|---|---|
| WeirdML | 9.8% | 12.4% |
| BigCodeBench Instruct | 39.4% | 46% |
| LMArena Coding | 1199 | 1254 |
| BigCodeBench Complete | 50.1% | 57.2% |
| HumanEval+ | 68.9% | 79.3% |
| CadEval | 12% | — |
| MBPP+ | 68.8% | — |

## Agentic & Tool Use

- Claude 3 Haiku: —
- GPT-4: —

| Benchmark | Claude 3 Haiku | GPT-4 |
|---|---|---|
| METR Time Horizons | — | 36.1% |

## Reasoning

- Claude 3 Haiku: 15.7 (#303)
- GPT-4: 17.3 (#285)

| Benchmark | Claude 3 Haiku | GPT-4 |
|---|---|---|
| LMArena Hard Prompts | 1174 | 1241 |
| DTBench | 50.1% | 62.7% |
| LMCA | 8.8% | 17.1% |
| Epoch Capabilities Index | 118.3 | 125.88 |
| ForecastBench | 53.2 | 57.8 |
| WinoGrande | 74.2% | 87.5% |
| Kagi LLM Benchmark | 34.2% | — |
| Chess Puzzles | — | 4% |
| Mystery Game Puzzles | — | 12% |
| BIG-Bench Hard | — | 75.1% |
| HellaSwag | — | 95.3% |

## Math

- Claude 3 Haiku: 9.2 (#313)
- GPT-4: 10.1 (#304)

| Benchmark | Claude 3 Haiku | GPT-4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.8% | 1.1% |
| LMArena Math | 1188 | 1269 |
| MATH Level 5 | 14.9% | 23% |
| GSM8K | — | 92% |

## Knowledge

- Claude 3 Haiku: 17.3 (#280)
- GPT-4: 18.4 (#277)

| Benchmark | Claude 3 Haiku | GPT-4 |
|---|---|---|
| GPQA Diamond | 36.3% | 35.7% |
| LMArena Expert | 1148 | 1211 |
| MMLU | 73.8% | 86.4% |
| Confabulations | 34.2% | — |
| TriviaQA | — | 84.8% |

## Multimodal

- Claude 3 Haiku: 23.1 (#123)
- GPT-4: —

| Benchmark | Claude 3 Haiku | GPT-4 |
|---|---|---|
| LMArena Vision | 950 | — |
| ScienceQA | 72% | — |

## Multilingual

- Claude 3 Haiku: 36.1 (#238)
- GPT-4: 40.8 (#210)

| Benchmark | Claude 3 Haiku | GPT-4 |
|---|---|---|
| LMArena Non-English | 1178 | 1246 |
| LMArena Chinese | 1155 | 1242 |
| LMArena French | 1194 | 1282 |
| LMArena German | 1174 | 1251 |
| LMArena Japanese | 1102 | 1209 |
| LMArena Korean | 1108 | 1184 |
| LMArena Russian | 1204 | 1251 |
| LMArena Spanish | 1165 | 1261 |

## Instruction Following

- Claude 3 Haiku: 61.5 (#242)
- GPT-4: 65.5 (#217)

| Benchmark | Claude 3 Haiku | GPT-4 |
|---|---|---|
| LMArena Instruction Following | 1174 | 1241 |

## Long Context

- Claude 3 Haiku: 36.1 (#232)
- GPT-4: 37.7 (#207)

| Benchmark | Claude 3 Haiku | GPT-4 |
|---|---|---|
| LMArena Longer Query | 1191 | 1244 |

## Writing & Preference

- Claude 3 Haiku: 29.8 (#286)
- GPT-4: 35.0 (#263)

| Benchmark | Claude 3 Haiku | GPT-4 |
|---|---|---|
| LMArena Text | 1195 | 1263 |
| LMArena Creative Writing | 1157 | 1244 |
| EQ-Bench Creative Writing | 717 | 752 |
| LMArena Multi-Turn | 1190 | 1257 |

## FAQ

### Is Claude 3 Haiku better than GPT-4?

GPT-4 is the stronger model overall, scoring 28.9 to 25.7 on the Noometry Index.

### Is Claude 3 Haiku or GPT-4 better for coding?

GPT-4 scores higher on coding benchmarks: 31.2 versus 26.1 in the Noometry coding category.

### How many benchmarks do Claude 3 Haiku and GPT-4 share?

31 benchmarks have published results for both models. Claude 3 Haiku has 37 scored results on Noometry and GPT-4 has 38.