# Claude 3 Haiku vs Llama 2-70B

> Claude 3 Haiku is the stronger model overall, scoring 25.7 to 24.1 on the Noometry Index.

- Canonical page: https://noometry.com/compare/claude-3-haiku-vs-llama-2-70b
- Last updated: 2026-10-10
- Shared benchmarks: 25

## Summary

- They share 25 benchmarks with published results for both. Claude 3 Haiku scores higher in 6 categories and Llama 2-70B in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Claude 3 Haiku leads 17.3 to 7.4.
- The biggest single-benchmark swing is MATH Level 5: 14.9% for Claude 3 Haiku and 3.3% for Llama 2-70B.
- Llama 2-70B has downloadable open weights; the other is API-only.

## Snapshot

| | Claude 3 Haiku | Llama 2-70B |
|---|---|---|
| Provider | Anthropic | Meta |
| Noometry Index | 25.7 | 24.1 |
| Rank | 333 | 343 |
| Context | — | — |
| Input $/M | — | — |
| Output $/M | — | — |
| Weights | Proprietary | Open |

## Coding

- Claude 3 Haiku: 26.1 (#320)
- Llama 2-70B: 30.8 (#281)

| Benchmark | Claude 3 Haiku | Llama 2-70B |
|---|---|---|
| LMArena Coding | 1199 | 1079 |
| WeirdML | 9.8% | — |
| BigCodeBench Instruct | 39.4% | — |
| BigCodeBench Complete | 50.1% | — |
| CadEval | 12% | — |
| HumanEval+ | 68.9% | — |
| MBPP+ | 68.8% | — |

## Reasoning

- Claude 3 Haiku: 15.7 (#303)
- Llama 2-70B: 13.7 (#324)

| Benchmark | Claude 3 Haiku | Llama 2-70B |
|---|---|---|
| LMArena Hard Prompts | 1174 | 1073 |
| DTBench | 50.1% | 41.6% |
| Epoch Capabilities Index | 118.3 | 113.59 |
| ForecastBench | 53.2 | 51.4 |
| WinoGrande | 74.2% | 80.2% |
| Kagi LLM Benchmark | 34.2% | — |
| LMCA | 8.8% | — |
| BIG-Bench Hard | — | 64.9% |
| CommonsenseQA 2.0 | — | 50% |
| HellaSwag | — | 85.3% |
| LAMBADA | — | 78.9% |
| PIQA | — | 82.8% |

## Math

- Claude 3 Haiku: 9.2 (#313)
- Llama 2-70B: 7.5 (#320)

| Benchmark | Claude 3 Haiku | Llama 2-70B |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.8% | 0% |
| LMArena Math | 1188 | 1091 |
| MATH Level 5 | 14.9% | 3.3% |
| GSM8K | — | 69.6% |

## Knowledge

- Claude 3 Haiku: 17.3 (#280)
- Llama 2-70B: 7.4 (#305)

| Benchmark | Claude 3 Haiku | Llama 2-70B |
|---|---|---|
| GPQA Diamond | 36.3% | 26.3% |
| LMArena Expert | 1148 | 1039 |
| MMLU | 73.8% | 69.9% |
| Confabulations | 34.2% | — |
| ARC (AI2) Challenge | — | 78.3% |
| BoolQ | — | 88.6% |
| OpenBookQA | — | 60.2% |
| TriviaQA | — | 87.6% |

## Multimodal

- Claude 3 Haiku: 23.1 (#123)
- Llama 2-70B: —

| Benchmark | Claude 3 Haiku | Llama 2-70B |
|---|---|---|
| LMArena Vision | 950 | — |
| ScienceQA | 72% | — |

## Multilingual

- Claude 3 Haiku: 36.1 (#238)
- Llama 2-70B: 27.9 (#269)

| Benchmark | Claude 3 Haiku | Llama 2-70B |
|---|---|---|
| LMArena Non-English | 1178 | 1045 |
| LMArena Chinese | 1155 | 994 |
| LMArena French | 1194 | 1089 |
| LMArena German | 1174 | 1041 |
| LMArena Japanese | 1102 | 927 |
| LMArena Korean | 1108 | 964 |
| LMArena Russian | 1204 | 1083 |
| LMArena Spanish | 1165 | 1142 |

## Instruction Following

- Claude 3 Haiku: 61.5 (#242)
- Llama 2-70B: 55.1 (#273)

| Benchmark | Claude 3 Haiku | Llama 2-70B |
|---|---|---|
| LMArena Instruction Following | 1174 | 1071 |

## Long Context

- Claude 3 Haiku: 36.1 (#232)
- Llama 2-70B: 32.2 (#265)

| Benchmark | Claude 3 Haiku | Llama 2-70B |
|---|---|---|
| LMArena Longer Query | 1191 | 1062 |

## Writing & Preference

- Claude 3 Haiku: 29.8 (#286)
- Llama 2-70B: 32.5 (#274)

| Benchmark | Claude 3 Haiku | Llama 2-70B |
|---|---|---|
| LMArena Text | 1195 | 1115 |
| LMArena Creative Writing | 1157 | 1075 |
| LMArena Multi-Turn | 1190 | 1088 |
| EQ-Bench Creative Writing | 717 | — |

## FAQ

### Is Claude 3 Haiku better than Llama 2-70B?

Claude 3 Haiku is the stronger model overall, scoring 25.7 to 24.1 on the Noometry Index.

### Is Claude 3 Haiku or Llama 2-70B better for coding?

Llama 2-70B scores higher on coding benchmarks: 30.8 versus 26.1 in the Noometry coding category.

### How many benchmarks do Claude 3 Haiku and Llama 2-70B share?

25 benchmarks have published results for both models. Claude 3 Haiku has 37 scored results on Noometry and Llama 2-70B has 35.