Model comparison
Llama 2-7B vs Qwen2.5-Coder-32B
Qwen2.5-Coder-32B is the stronger model overall, scoring 33.4 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 Qwen2.5-Coder-32B in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in multilingual, where Qwen2.5-Coder-32B leads 37.8 to 23.8.
Side by side
| Llama 2-7B | Qwen2.5-Coder-32B | |
|---|---|---|
| Provider | Meta | Alibaba (Qwen) |
| Noometry Index | 29.1 | 33.4 |
| Released | 2023-07-18 | 2024-09-18 |
| Weights | Open | Open |
| Context window | — | 33K |
| Max output | — | 29K |
| Input $ / M tokens | — | $0.66 |
| Output $ / M tokens | — | $1 |
| Results tracked | 29 | 31 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Llama 2-7B leads
Llama 2-7B: 29.2 (#307), Qwen2.5-Coder-32B: 22.6 (#333)
| Benchmark | Llama 2-7B | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Coding | 1002 | 1276 |
| SWE-bench Verified (bash only) | — | 9% |
| Aider Polyglot | — | 16.4% |
| BigCodeBench Instruct | — | 49% |
| LiveBench Coding | — | 56.9% |
| BigCodeBench Complete | — | 58% |
| HumanEval+ | — | 87.2% |
| MBPP+ | — | 77% |
Reasoning Qwen2.5-Coder-32B leads
Llama 2-7B: 15.7 (#312), Qwen2.5-Coder-32B: 21.2 (#225)
| Benchmark | Llama 2-7B | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Hard Prompts | 1009 | 1251 |
| Epoch Capabilities Index | 99.06 | 119.49 |
| HellaSwag | 77.2% | 83% |
| WinoGrande | 69.2% | 80.8% |
| Chess Puzzles | 0% | — |
| LiveBench Reasoning | — | 42.1% |
| LiveBench Data Analysis | — | 49.9% |
| BIG-Bench Hard | 39.2% | — |
| LAMBADA | 73.3% | — |
| LiveBench | — | 46.2% |
| PIQA | 78.8% | — |
Math Qwen2.5-Coder-32B leads
Llama 2-7B: 30.7 (#233), Qwen2.5-Coder-32B: 33.3 (#204)
| Benchmark | Llama 2-7B | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Math | 1042 | 1251 |
| GSM8K | 16.7% | 93% |
| LiveBench Math | — | 46.6% |
Knowledge Qwen2.5-Coder-32B leads
Llama 2-7B: 28.2 (#248), Qwen2.5-Coder-32B: 33.4 (#203)
| Benchmark | Llama 2-7B | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Expert | 1036 | 1221 |
| ARC (AI2) Challenge | 45.9% | 70.5% |
| MMLU | 45.8% | 79.1% |
| BoolQ | 77.9% | — |
| OpenBookQA | 58.6% | — |
| TriviaQA | 73.7% | — |
Multimodal Not comparable
Llama 2-7B: —, Qwen2.5-Coder-32B: —
| Benchmark | Llama 2-7B | Qwen2.5-Coder-32B |
|---|---|---|
| ScienceQA | 43.1% | — |
Multilingual Qwen2.5-Coder-32B leads
Llama 2-7B: 23.8 (#293), Qwen2.5-Coder-32B: 37.8 (#235)
| Benchmark | Llama 2-7B | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Non-English | 973 | 1205 |
| LMArena Chinese | 973 | 1222 |
| LMArena Russian | 995 | 1228 |
| LMArena French | 970 | — |
| LMArena German | 978 | — |
| LMArena Spanish | 1007 | — |
Instruction Following Qwen2.5-Coder-32B leads
Llama 2-7B: 50.8 (#298), Qwen2.5-Coder-32B: 61.4 (#245)
| Benchmark | Llama 2-7B | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Instruction Following | 1006 | 1223 |
| LiveBench Instruction Following | — | 58.7% |
Long Context Qwen2.5-Coder-32B leads
Llama 2-7B: 30.4 (#287), Qwen2.5-Coder-32B: 38.0 (#208)
| Benchmark | Llama 2-7B | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Longer Query | 999 | 1251 |
Writing & Preference Qwen2.5-Coder-32B leads
Llama 2-7B: 28.0 (#298), Qwen2.5-Coder-32B: 41.6 (#240)
| Benchmark | Llama 2-7B | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Text | 1053 | 1230 |
| LMArena Creative Writing | 1033 | 1174 |
| LMArena Multi-Turn | 1029 | 1222 |
| LiveBench Language | — | 23.3% |
Frequently asked questions
Is Llama 2-7B better than Qwen2.5-Coder-32B?
Qwen2.5-Coder-32B is the stronger model overall, scoring 33.4 to 29.1 on the Noometry Index.
Is Llama 2-7B or Qwen2.5-Coder-32B better for coding?
Llama 2-7B scores higher on coding benchmarks: 29.2 versus 22.6 in the Noometry coding category.
How many benchmarks do Llama 2-7B and Qwen2.5-Coder-32B share?
18 benchmarks have published results for both models. Llama 2-7B has 29 scored results on Noometry and Qwen2.5-Coder-32B has 31.