Model comparison
Qwen1.5-110B vs Qwen2-72B
Qwen1.5-110B is the stronger model overall, scoring 34.2 to 30.0 on the Noometry Index.
Last verified . 19 shared benchmarks.
Alibaba (Qwen)Summary
- They share 19 benchmarks with published results for both. Qwen1.5-110B scores higher in 3 categories and Qwen2-72B in 5 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where Qwen1.5-110B leads 31.2 to 21.2.
- The biggest single-benchmark swing is BigCodeBench Complete: 44.4% for Qwen1.5-110B and 54% for Qwen2-72B.
Side by side
| Qwen1.5-110B | Qwen2-72B | |
|---|---|---|
| Provider | Alibaba (Qwen) | Alibaba (Qwen) |
| Noometry Index | 34.2 | 30.0 |
| Released | 2024-04-25 | 2024-06-07 |
| Weights | Open | Open |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 20 | 26 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen1.5-110B leads
Qwen1.5-110B: 33.0 (#264), Qwen2-72B: 29.1 (#310)
| Benchmark | Qwen1.5-110B | Qwen2-72B |
|---|---|---|
| BigCodeBench Instruct | 35% | 38.5% |
| LMArena Coding | 1184 | 1196 |
| BigCodeBench Complete | 44.4% | 54% |
| WeirdML | — | 11.3% |
Agentic & Tool Use Not comparable
Qwen1.5-110B: —, Qwen2-72B: 17.0 (#146)
| Benchmark | Qwen1.5-110B | Qwen2-72B |
|---|---|---|
| TheAgentCompany | — | 1.1% |
| METR Time Horizons | — | 29.9% |
Reasoning Too close to call
Qwen1.5-110B: 22.7 (#189), Qwen2-72B: 23.2 (#181)
| Benchmark | Qwen1.5-110B | Qwen2-72B |
|---|---|---|
| LMArena Hard Prompts | 1168 | 1191 |
| Epoch Capabilities Index | — | 125.28 |
| ForecastBench | 57.7 | — |
Math Qwen1.5-110B leads
Qwen1.5-110B: 33.7 (#201), Qwen2-72B: 30.2 (#236)
| Benchmark | Qwen1.5-110B | Qwen2-72B |
|---|---|---|
| LMArena Math | 1185 | 1235 |
| MATH Level 5 | — | 39.1% |
Knowledge Qwen1.5-110B leads
Qwen1.5-110B: 31.2 (#219), Qwen2-72B: 21.2 (#275)
| Benchmark | Qwen1.5-110B | Qwen2-72B |
|---|---|---|
| LMArena Expert | 1144 | 1171 |
| GPQA Diamond | — | 40.8% |
| MMLU | — | 82.4% |
Multilingual Qwen2-72B leads
Qwen1.5-110B: 33.6 (#250), Qwen2-72B: 35.9 (#244)
| Benchmark | Qwen1.5-110B | Qwen2-72B |
|---|---|---|
| LMArena Non-English | 1142 | 1176 |
| LMArena Chinese | 1206 | 1240 |
| LMArena French | 1151 | 1170 |
| LMArena German | 1123 | 1151 |
| LMArena Japanese | 1074 | 1111 |
| LMArena Korean | 1044 | 1083 |
| LMArena Russian | 1118 | 1169 |
| LMArena Spanish | 1142 | 1169 |
Instruction Following Qwen2-72B leads
Qwen1.5-110B: 60.3 (#252), Qwen2-72B: 61.7 (#241)
| Benchmark | Qwen1.5-110B | Qwen2-72B |
|---|---|---|
| LMArena Instruction Following | 1158 | 1181 |
Long Context Qwen2-72B leads
Qwen1.5-110B: 35.1 (#242), Qwen2-72B: 36.1 (#235)
| Benchmark | Qwen1.5-110B | Qwen2-72B |
|---|---|---|
| LMArena Longer Query | 1157 | 1192 |
Writing & Preference Qwen2-72B leads
Qwen1.5-110B: 38.0 (#255), Qwen2-72B: 40.8 (#241)
| Benchmark | Qwen1.5-110B | Qwen2-72B |
|---|---|---|
| LMArena Text | 1175 | 1203 |
| LMArena Creative Writing | 1148 | 1181 |
| LMArena Multi-Turn | 1160 | 1196 |
Frequently asked questions
Is Qwen1.5-110B better than Qwen2-72B?
Qwen1.5-110B is the stronger model overall, scoring 34.2 to 30.0 on the Noometry Index.
Is Qwen1.5-110B or Qwen2-72B better for coding?
Qwen1.5-110B scores higher on coding benchmarks: 33.0 versus 29.1 in the Noometry coding category.
How many benchmarks do Qwen1.5-110B and Qwen2-72B share?
19 benchmarks have published results for both models. Qwen1.5-110B has 20 scored results on Noometry and Qwen2-72B has 26.