Model comparison
Qwen-14B vs Qwen2-72B
Qwen-14B is the stronger model overall, scoring 30.8 to 29.5 on the Noometry Index.
Last verified . 12 shared benchmarks.
Alibaba (Qwen)Summary
- They share 12 benchmarks with published results for both. Qwen-14B scores higher in 2 categories and Qwen2-72B in 5 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Qwen2-72B leads 41.0 to 27.8.
Side by side
| Qwen-14B | Qwen2-72B | |
|---|---|---|
| Provider | Alibaba (Qwen) | Alibaba (Qwen) |
| Noometry Index | 30.8 | 29.5 |
| Released | 2023-09-24 | 2024-06-07 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 18 | 26 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen-14B leads
Qwen-14B: 30.6 (#283), Qwen2-72B: 28.7 (#303)
| Benchmark | Qwen-14B | Qwen2-72B |
|---|---|---|
| LMArena Coding | 1072 | 1196 |
| WeirdML | — | 11.3% |
| BigCodeBench Instruct | — | 38.5% |
| BigCodeBench Complete | — | 54% |
Agentic & Tool Use Not comparable
Qwen-14B: —, Qwen2-72B: 16.9 (#138)
| Benchmark | Qwen-14B | Qwen2-72B |
|---|---|---|
| TheAgentCompany | — | 1.1% |
| METR Time Horizons | — | 29.9% |
Reasoning Qwen2-72B leads
Qwen-14B: 18.7 (#255), Qwen2-72B: 22.2 (#184)
| Benchmark | Qwen-14B | Qwen2-72B |
|---|---|---|
| LMArena Hard Prompts | 1027 | 1191 |
| Epoch Capabilities Index | 112.8 | 125.27 |
| BIG-Bench Hard | 55% | — |
| LAMBADA | 71.1% | — |
| PIQA | 79.9% | — |
Math Too close to call
Qwen-14B: 28.7 (#224), Qwen2-72B: 27.9 (#232)
| Benchmark | Qwen-14B | Qwen2-72B |
|---|---|---|
| LMArena Math | 1068 | 1235 |
| MATH Level 5 | — | 39.1% |
| GSM8K | 61.3% | — |
Knowledge Not comparable
Qwen-14B: —, Qwen2-72B: 21.2 (#270)
| Benchmark | Qwen-14B | Qwen2-72B |
|---|---|---|
| MMLU | 66.3% | 82.4% |
| GPQA Diamond | — | 40.8% |
| LMArena Expert | — | 1171 |
| ARC (AI2) Challenge | 84.4% | — |
| BoolQ | 86.2% | — |
Multilingual Qwen2-72B leads
Qwen-14B: 27.7 (#270), Qwen2-72B: 36.0 (#239)
| Benchmark | Qwen-14B | Qwen2-72B |
|---|---|---|
| LMArena Non-English | 1042 | 1176 |
| LMArena Chinese | 1077 | 1240 |
| LMArena French | — | 1169 |
| LMArena German | — | 1151 |
| LMArena Japanese | — | 1111 |
| LMArena Korean | — | 1083 |
| LMArena Russian | — | 1169 |
| LMArena Spanish | — | 1168 |
Instruction Following Qwen2-72B leads
Qwen-14B: 52.6 (#284), Qwen2-72B: 62.0 (#236)
| Benchmark | Qwen-14B | Qwen2-72B |
|---|---|---|
| LMArena Instruction Following | 1031 | 1181 |
Long Context Qwen2-72B leads
Qwen-14B: 31.3 (#276), Qwen2-72B: 36.1 (#230)
| Benchmark | Qwen-14B | Qwen2-72B |
|---|---|---|
| LMArena Longer Query | 1029 | 1193 |
Writing & Preference Qwen2-72B leads
Qwen-14B: 27.8 (#294), Qwen2-72B: 41.0 (#236)
| Benchmark | Qwen-14B | Qwen2-72B |
|---|---|---|
| LMArena Text | 1051 | 1203 |
| LMArena Creative Writing | 1028 | 1181 |
| LMArena Multi-Turn | 1022 | 1195 |
Frequently asked questions
Is Qwen-14B better than Qwen2-72B?
Qwen-14B is the stronger model overall, scoring 30.8 to 29.5 on the Noometry Index.
Is Qwen-14B or Qwen2-72B better for coding?
Qwen-14B scores higher on coding benchmarks: 30.6 versus 28.7 in the Noometry coding category.
How many benchmarks do Qwen-14B and Qwen2-72B share?
12 benchmarks have published results for both models. Qwen-14B has 18 scored results on Noometry and Qwen2-72B has 26.