# Hy3 vs Llama 2-70B

> Hy3 is the stronger model overall, scoring 43.6 to 24.1 on the Noometry Index.

- Canonical page: https://noometry.com/compare/hy3-vs-llama-2-70b
- Last updated: 2026-10-10
- Shared benchmarks: 15

## Summary

- They share 15 benchmarks with published results for both. Hy3 scores higher in 8 categories and Llama 2-70B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Hy3 leads 40.2 to 7.4.

## Snapshot

| | Hy3 | Llama 2-70B |
|---|---|---|
| Provider | Tencent | Meta |
| Noometry Index | 43.6 | 24.1 |
| Rank | 79 | 343 |
| Context | 262K | — |
| Input $/M | $0.13 | — |
| Output $/M | $0.53 | — |
| Weights | Open | Open |

## Coding

- Hy3: 46.6 (#60)
- Llama 2-70B: 30.8 (#281)

| Benchmark | Hy3 | Llama 2-70B |
|---|---|---|
| LMArena Coding | 1465 | 1079 |
| LMArena WebDev | 1509 | — |

## Reasoning

- Hy3: 25.0 (#134)
- Llama 2-70B: 13.7 (#324)

| Benchmark | Hy3 | Llama 2-70B |
|---|---|---|
| LMArena Hard Prompts | 1449 | 1073 |
| NYT Connections (extended) | 41.2% | — |
| DTBench | — | 41.6% |
| BIG-Bench Hard | — | 64.9% |
| CommonsenseQA 2.0 | — | 50% |
| Epoch Capabilities Index | — | 113.59 |
| ForecastBench | — | 51.4 |
| HellaSwag | — | 85.3% |
| LAMBADA | — | 78.9% |
| PIQA | — | 82.8% |
| WinoGrande | — | 80.2% |

## Math

- Hy3: 37.4 (#98)
- Llama 2-70B: 7.5 (#320)

| Benchmark | Hy3 | Llama 2-70B |
|---|---|---|
| LMArena Math | 1475 | 1091 |
| OTIS Mock AIME 2024-2025 | — | 0% |
| MATH Level 5 | — | 3.3% |
| GSM8K | — | 69.6% |

## Knowledge

- Hy3: 40.2 (#108)
- Llama 2-70B: 7.4 (#305)

| Benchmark | Hy3 | Llama 2-70B |
|---|---|---|
| LMArena Expert | 1461 | 1039 |
| GPQA Diamond | — | 26.3% |
| ARC (AI2) Challenge | — | 78.3% |
| BoolQ | — | 88.6% |
| MMLU | — | 69.9% |
| OpenBookQA | — | 60.2% |
| TriviaQA | — | 87.6% |

## Multilingual

- Hy3: 53.7 (#60)
- Llama 2-70B: 27.9 (#269)

| Benchmark | Hy3 | Llama 2-70B |
|---|---|---|
| LMArena Non-English | 1427 | 1045 |
| LMArena Chinese | 1496 | 994 |
| LMArena French | 1461 | 1089 |
| LMArena Korean | 1390 | 964 |
| LMArena Russian | 1435 | 1083 |
| LMArena Spanish | 1448 | 1142 |
| LMArena German | — | 1041 |
| LMArena Japanese | — | 927 |

## Instruction Following

- Hy3: 75.3 (#68)
- Llama 2-70B: 55.1 (#273)

| Benchmark | Hy3 | Llama 2-70B |
|---|---|---|
| LMArena Instruction Following | 1425 | 1071 |

## Long Context

- Hy3: 44.1 (#70)
- Llama 2-70B: 32.2 (#265)

| Benchmark | Hy3 | Llama 2-70B |
|---|---|---|
| LMArena Longer Query | 1443 | 1062 |

## Writing & Preference

- Hy3: 62.5 (#75)
- Llama 2-70B: 32.5 (#274)

| Benchmark | Hy3 | Llama 2-70B |
|---|---|---|
| LMArena Text | 1442 | 1115 |
| LMArena Creative Writing | 1399 | 1075 |
| LMArena Multi-Turn | 1436 | 1088 |

## FAQ

### Is Hy3 better than Llama 2-70B?

Hy3 is the stronger model overall, scoring 43.6 to 24.1 on the Noometry Index.

### Is Hy3 or Llama 2-70B better for coding?

Hy3 scores higher on coding benchmarks: 46.6 versus 30.8 in the Noometry coding category.

### How many benchmarks do Hy3 and Llama 2-70B share?

15 benchmarks have published results for both models. Hy3 has 17 scored results on Noometry and Llama 2-70B has 35.