Model comparison
Llama 2-7B vs Qwen-14B
Qwen-14B is the stronger model overall, scoring 31.4 to 29.1 on the Noometry Index.
Last verified . 18 shared benchmarks.
Meta
Alibaba (Qwen)Summary
- They share 18 benchmarks with published results for both. Llama 2-7B scores higher in 1 category and Qwen-14B in 6 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Qwen-14B leads 19.6 to 15.7.
Side by side
| Llama 2-7B | Qwen-14B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 29.1 | 31.4 |
| Released | 2023-07-18 | 2023-09-24 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 29 | 18 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen-14B leads
Llama 2-7B: 29.2 (#307), Qwen-14B: 31.2 (#288)
| Benchmark | Llama 2-7B | Qwen-14B |
|---|---|---|
| LMArena Coding | 1002 | 1071 |
Reasoning Qwen-14B leads
Llama 2-7B: 15.7 (#312), Qwen-14B: 19.6 (#257)
| Benchmark | Llama 2-7B | Qwen-14B |
|---|---|---|
| LMArena Hard Prompts | 1009 | 1027 |
| BIG-Bench Hard | 39.2% | 55% |
| Epoch Capabilities Index | 99.06 | 113.03 |
| LAMBADA | 73.3% | 71.1% |
| PIQA | 78.8% | 79.9% |
| Chess Puzzles | 0% | — |
| HellaSwag | 77.2% | — |
| WinoGrande | 69.2% | — |
Math Too close to call
Llama 2-7B: 30.7 (#233), Qwen-14B: 31.2 (#227)
| Benchmark | Llama 2-7B | Qwen-14B |
|---|---|---|
| LMArena Math | 1042 | 1068 |
| GSM8K | 16.7% | 61.3% |
Knowledge Not comparable
Llama 2-7B: 28.2 (#248), Qwen-14B: —
| Benchmark | Llama 2-7B | Qwen-14B |
|---|---|---|
| ARC (AI2) Challenge | 45.9% | 84.4% |
| BoolQ | 77.9% | 86.2% |
| MMLU | 45.8% | 66.3% |
| LMArena Expert | 1036 | — |
| OpenBookQA | 58.6% | — |
| TriviaQA | 73.7% | — |
Multimodal Not comparable
Llama 2-7B: —, Qwen-14B: —
| Benchmark | Llama 2-7B | Qwen-14B |
|---|---|---|
| ScienceQA | 43.1% | — |
Multilingual Qwen-14B leads
Llama 2-7B: 23.8 (#293), Qwen-14B: 27.5 (#275)
| Benchmark | Llama 2-7B | Qwen-14B |
|---|---|---|
| LMArena Non-English | 973 | 1041 |
| LMArena Chinese | 973 | 1077 |
| LMArena French | 970 | — |
| LMArena German | 978 | — |
| LMArena Russian | 995 | — |
| LMArena Spanish | 1007 | — |
Instruction Following Qwen-14B leads
Llama 2-7B: 50.8 (#298), Qwen-14B: 52.4 (#289)
| Benchmark | Llama 2-7B | Qwen-14B |
|---|---|---|
| LMArena Instruction Following | 1006 | 1031 |
Long Context Too close to call
Llama 2-7B: 30.4 (#287), Qwen-14B: 31.3 (#280)
| Benchmark | Llama 2-7B | Qwen-14B |
|---|---|---|
| LMArena Longer Query | 999 | 1028 |
Writing & Preference Too close to call
Llama 2-7B: 28.0 (#298), Qwen-14B: 27.6 (#299)
| Benchmark | Llama 2-7B | Qwen-14B |
|---|---|---|
| LMArena Text | 1053 | 1051 |
| LMArena Creative Writing | 1033 | 1028 |
| LMArena Multi-Turn | 1029 | 1022 |
Frequently asked questions
Is Llama 2-7B better than Qwen-14B?
Qwen-14B is the stronger model overall, scoring 31.4 to 29.1 on the Noometry Index.
Is Llama 2-7B or Qwen-14B better for coding?
Qwen-14B scores higher on coding benchmarks: 31.2 versus 29.2 in the Noometry coding category.
How many benchmarks do Llama 2-7B and Qwen-14B share?
18 benchmarks have published results for both models. Llama 2-7B has 29 scored results on Noometry and Qwen-14B has 18.