Model comparison
Hy3 vs Llama 2-7B
Hy3 is the stronger model overall, scoring 44.2 to 29.1 on the Noometry Index.
Last verified . 15 shared benchmarks.
Tencent
MetaSummary
- They share 15 benchmarks with published results for both. Hy3 scores higher in 8 categories and Llama 2-7B in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Hy3 leads 62.2 to 28.0.
Side by side
| Hy3 | Llama 2-7B | |
|---|---|---|
| Provider | Tencent | Meta |
| Noometry Index | 44.2 | 29.1 |
| Released | 2026-07-06 | 2023-07-18 |
| Weights | Open | Open |
| Context window | 262K | — |
| Max output | 128K | — |
| Input $ / M tokens | $0.0825 | — |
| Output $ / M tokens | $0.33 | — |
| Results tracked | 19 | 29 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Hy3 leads
Hy3: 46.8 (#63), Llama 2-7B: 29.2 (#307)
| Benchmark | Hy3 | Llama 2-7B |
|---|---|---|
| LMArena Coding | 1464 | 1002 |
| LMArena WebDev | 1508 | — |
Reasoning Hy3 leads
Hy3: 26.1 (#136), Llama 2-7B: 15.7 (#312)
| Benchmark | Hy3 | Llama 2-7B |
|---|---|---|
| LMArena Hard Prompts | 1447 | 1009 |
| NYT Connections (extended) | 41.2% | — |
| Chess Puzzles | — | 0% |
| BIG-Bench Hard | — | 39.2% |
| Epoch Capabilities Index | — | 99.06 |
| HellaSwag | — | 77.2% |
| LAMBADA | — | 73.3% |
| PIQA | — | 78.8% |
| WinoGrande | — | 69.2% |
Math Hy3 leads
Hy3: 40.1 (#93), Llama 2-7B: 30.7 (#233)
| Benchmark | Hy3 | Llama 2-7B |
|---|---|---|
| LMArena Math | 1475 | 1042 |
| GSM8K | — | 16.7% |
Knowledge Hy3 leads
Hy3: 40.8 (#114), Llama 2-7B: 28.2 (#248)
| Benchmark | Hy3 | Llama 2-7B |
|---|---|---|
| LMArena Expert | 1460 | 1036 |
| ARC (AI2) Challenge | — | 45.9% |
| BoolQ | — | 77.9% |
| MMLU | — | 45.8% |
| OpenBookQA | — | 58.6% |
| TriviaQA | — | 73.7% |
Multimodal Not comparable
Hy3: —, Llama 2-7B: —
| Benchmark | Hy3 | Llama 2-7B |
|---|---|---|
| ScienceQA | — | 43.1% |
Multilingual Hy3 leads
Hy3: 53.5 (#65), Llama 2-7B: 23.8 (#293)
| Benchmark | Hy3 | Llama 2-7B |
|---|---|---|
| LMArena Non-English | 1426 | 973 |
| LMArena Chinese | 1493 | 973 |
| LMArena French | 1461 | 970 |
| LMArena German | 1439 | 978 |
| LMArena Russian | 1432 | 995 |
| LMArena Spanish | 1456 | 1007 |
| LMArena Japanese | 1392 | — |
| LMArena Korean | 1395 | — |
Instruction Following Hy3 leads
Hy3: 75.1 (#70), Llama 2-7B: 50.8 (#298)
| Benchmark | Hy3 | Llama 2-7B |
|---|---|---|
| LMArena Instruction Following | 1426 | 1006 |
Long Context Hy3 leads
Hy3: 44.1 (#75), Llama 2-7B: 30.4 (#287)
| Benchmark | Hy3 | Llama 2-7B |
|---|---|---|
| LMArena Longer Query | 1442 | 999 |
Writing & Preference Hy3 leads
Hy3: 62.2 (#81), Llama 2-7B: 28.0 (#298)
| Benchmark | Hy3 | Llama 2-7B |
|---|---|---|
| LMArena Text | 1439 | 1053 |
| LMArena Creative Writing | 1402 | 1033 |
| LMArena Multi-Turn | 1436 | 1029 |
Frequently asked questions
Is Hy3 better than Llama 2-7B?
Hy3 is the stronger model overall, scoring 44.2 to 29.1 on the Noometry Index.
Is Hy3 or Llama 2-7B better for coding?
Hy3 scores higher on coding benchmarks: 46.8 versus 29.2 in the Noometry coding category.
How many benchmarks do Hy3 and Llama 2-7B share?
15 benchmarks have published results for both models. Hy3 has 19 scored results on Noometry and Llama 2-7B has 29.