Model comparison
Qwen2.5 72B Instruct vs Qwen2-72B
Qwen2.5 72B Instruct is the stronger model overall, scoring 31.9 to 30.0 on the Noometry Index.
Last verified . 26 shared benchmarks.
Alibaba (Qwen)Summary
- They share 26 benchmarks with published results for both. Qwen2.5 72B Instruct scores higher in 7 categories and Qwen2-72B in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen2-72B leads 30.2 to 19.3.
- The biggest single-benchmark swing is MATH Level 5: 63.2% for Qwen2.5 72B Instruct and 39.1% for Qwen2-72B.
Side by side
| Qwen2.5 72B Instruct | Qwen2-72B | |
|---|---|---|
| Provider | Alibaba (Qwen) | Alibaba (Qwen) |
| Noometry Index | 31.9 | 30.0 |
| Released | 2024-09 | 2024-06-07 |
| Weights | Open | Open |
| Context window | 131K | — |
| Max output | 8K | — |
| Input $ / M tokens | $1.40 | — |
| Output $ / M tokens | $5.60 | — |
| Results tracked | 43 | 26 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen2.5 72B Instruct leads
Qwen2.5 72B Instruct: 33.2 (#260), Qwen2-72B: 29.1 (#310)
| Benchmark | Qwen2.5 72B Instruct | Qwen2-72B |
|---|---|---|
| WeirdML | 16% | 11.3% |
| BigCodeBench Instruct | 45.8% | 38.5% |
| LMArena Coding | 1292 | 1196 |
| BigCodeBench Complete | 55.9% | 54% |
Agentic & Tool Use Qwen2.5 72B Instruct leads
Qwen2.5 72B Instruct: 22.1 (#133), Qwen2-72B: 17.0 (#146)
| Benchmark | Qwen2.5 72B Instruct | Qwen2-72B |
|---|---|---|
| TheAgentCompany | 5.7% | 1.1% |
| METR Time Horizons | 35.8% | 29.9% |
| BALROG | 16.2% | — |
Reasoning Too close to call
Qwen2.5 72B Instruct: 22.3 (#199), Qwen2-72B: 23.2 (#181)
| Benchmark | Qwen2.5 72B Instruct | Qwen2-72B |
|---|---|---|
| LMArena Hard Prompts | 1271 | 1191 |
| Epoch Capabilities Index | 129 | 125.28 |
| DTBench | 62.9% | — |
| LMCA | 13.4% | — |
| BIG-Bench Hard | 79.8% | — |
| ForecastBench | 57.5 | — |
| HellaSwag | 84.8% | — |
| PIQA | 82.6% | — |
| WinoGrande | 82.3% | — |
Math Qwen2-72B leads
Qwen2.5 72B Instruct: 19.3 (#287), Qwen2-72B: 30.2 (#236)
| Benchmark | Qwen2.5 72B Instruct | Qwen2-72B |
|---|---|---|
| LMArena Math | 1283 | 1235 |
| MATH Level 5 | 63.2% | 39.1% |
| OTIS Mock AIME 2024-2025 | 8.1% | — |
| Omni-MATH | 33% | — |
Knowledge Qwen2.5 72B Instruct leads
Qwen2.5 72B Instruct: 27.0 (#253), Qwen2-72B: 21.2 (#275)
| Benchmark | Qwen2.5 72B Instruct | Qwen2-72B |
|---|---|---|
| GPQA Diamond | 49.1% | 40.8% |
| LMArena Expert | 1245 | 1171 |
| MMLU | 85.3% | 82.4% |
| MMLU-Pro | 63.1% | — |
| Confabulations | 19.1% | — |
| GPQA (HELM) | 42.6% | — |
| ARC (AI2) Challenge | 94.5% | — |
| TriviaQA | 71.9% | — |
Multilingual Qwen2.5 72B Instruct leads
Qwen2.5 72B Instruct: 41.0 (#213), Qwen2-72B: 35.9 (#244)
| Benchmark | Qwen2.5 72B Instruct | Qwen2-72B |
|---|---|---|
| LMArena Non-English | 1252 | 1176 |
| LMArena Chinese | 1272 | 1240 |
| LMArena French | 1280 | 1170 |
| LMArena German | 1234 | 1151 |
| LMArena Japanese | 1180 | 1111 |
| LMArena Korean | 1188 | 1083 |
| LMArena Russian | 1264 | 1169 |
| LMArena Spanish | 1256 | 1169 |
Instruction Following Qwen2.5 72B Instruct leads
Qwen2.5 72B Instruct: 65.5 (#221), Qwen2-72B: 61.7 (#241)
| Benchmark | Qwen2.5 72B Instruct | Qwen2-72B |
|---|---|---|
| LMArena Instruction Following | 1254 | 1181 |
| IFEval | 80.6% | — |
Long Context Qwen2.5 72B Instruct leads
Qwen2.5 72B Instruct: 38.9 (#188), Qwen2-72B: 36.1 (#235)
| Benchmark | Qwen2.5 72B Instruct | Qwen2-72B |
|---|---|---|
| LMArena Longer Query | 1282 | 1192 |
Writing & Preference Qwen2.5 72B Instruct leads
Qwen2.5 72B Instruct: 46.7 (#215), Qwen2-72B: 40.8 (#241)
| Benchmark | Qwen2.5 72B Instruct | Qwen2-72B |
|---|---|---|
| LMArena Text | 1269 | 1203 |
| LMArena Creative Writing | 1221 | 1181 |
| LMArena Multi-Turn | 1272 | 1196 |
| WildBench | 80.2% | — |
Frequently asked questions
Is Qwen2.5 72B Instruct better than Qwen2-72B?
Qwen2.5 72B Instruct is the stronger model overall, scoring 31.9 to 30.0 on the Noometry Index.
Is Qwen2.5 72B Instruct or Qwen2-72B better for coding?
Qwen2.5 72B Instruct scores higher on coding benchmarks: 33.2 versus 29.1 in the Noometry coding category.
How many benchmarks do Qwen2.5 72B Instruct and Qwen2-72B share?
26 benchmarks have published results for both models. Qwen2.5 72B Instruct has 43 scored results on Noometry and Qwen2-72B has 26.