Model comparison
Qwen2.5 72B Instruct vs Qwen2.5-Coder-32B
Qwen2.5-Coder-32B is the stronger model overall, scoring 33.4 to 31.9 on the Noometry Index.
Last verified . 19 shared benchmarks.
Alibaba (Qwen)Summary
- They share 19 benchmarks with published results for both. Qwen2.5 72B Instruct scores higher in 6 categories and Qwen2.5-Coder-32B in 2 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen2.5-Coder-32B leads 33.3 to 19.3.
- Qwen2.5-Coder-32B is cheaper at $0.66 / $1 per million input/output tokens, against $1.40 / $5.60 for Qwen2.5 72B Instruct.
- Qwen2.5 72B Instruct accepts more context: 131K tokens versus 33K.
Side by side
| Qwen2.5 72B Instruct | Qwen2.5-Coder-32B | |
|---|---|---|
| Provider | Alibaba (Qwen) | Alibaba (Qwen) |
| Noometry Index | 31.9 | 33.4 |
| Released | 2024-09 | 2024-09-18 |
| Weights | Open | Open |
| Context window | 131K | 33K |
| Max output | 8K | 29K |
| Input $ / M tokens | $1.40 | $0.66 |
| Output $ / M tokens | $5.60 | $1 |
| Results tracked | 43 | 31 |
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), Qwen2.5-Coder-32B: 22.6 (#333)
| Benchmark | Qwen2.5 72B Instruct | Qwen2.5-Coder-32B |
|---|---|---|
| BigCodeBench Instruct | 45.8% | 49% |
| LMArena Coding | 1292 | 1276 |
| BigCodeBench Complete | 55.9% | 58% |
| SWE-bench Verified (bash only) | — | 9% |
| Aider Polyglot | — | 16.4% |
| WeirdML | 16% | — |
| LiveBench Coding | — | 56.9% |
| HumanEval+ | — | 87.2% |
| MBPP+ | — | 77% |
Agentic & Tool Use Not comparable
Qwen2.5 72B Instruct: 22.1 (#133), Qwen2.5-Coder-32B: —
| Benchmark | Qwen2.5 72B Instruct | Qwen2.5-Coder-32B |
|---|---|---|
| TheAgentCompany | 5.7% | — |
| BALROG | 16.2% | — |
| METR Time Horizons | 35.8% | — |
Reasoning Qwen2.5 72B Instruct leads
Qwen2.5 72B Instruct: 22.3 (#199), Qwen2.5-Coder-32B: 21.2 (#225)
| Benchmark | Qwen2.5 72B Instruct | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Hard Prompts | 1271 | 1251 |
| Epoch Capabilities Index | 129 | 119.49 |
| HellaSwag | 84.8% | 83% |
| WinoGrande | 82.3% | 80.8% |
| LiveBench Reasoning | — | 42.1% |
| DTBench | 62.9% | — |
| LiveBench Data Analysis | — | 49.9% |
| LMCA | 13.4% | — |
| BIG-Bench Hard | 79.8% | — |
| ForecastBench | 57.5 | — |
| LiveBench | — | 46.2% |
| PIQA | 82.6% | — |
Math Qwen2.5-Coder-32B leads
Qwen2.5 72B Instruct: 19.3 (#287), Qwen2.5-Coder-32B: 33.3 (#204)
| Benchmark | Qwen2.5 72B Instruct | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Math | 1283 | 1251 |
| OTIS Mock AIME 2024-2025 | 8.1% | — |
| Omni-MATH | 33% | — |
| LiveBench Math | — | 46.6% |
| MATH Level 5 | 63.2% | — |
| GSM8K | — | 93% |
Knowledge Qwen2.5-Coder-32B leads
Qwen2.5 72B Instruct: 27.0 (#253), Qwen2.5-Coder-32B: 33.4 (#203)
| Benchmark | Qwen2.5 72B Instruct | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Expert | 1245 | 1221 |
| ARC (AI2) Challenge | 94.5% | 70.5% |
| MMLU | 85.3% | 79.1% |
| GPQA Diamond | 49.1% | — |
| MMLU-Pro | 63.1% | — |
| Confabulations | 19.1% | — |
| GPQA (HELM) | 42.6% | — |
| TriviaQA | 71.9% | — |
Multilingual Qwen2.5 72B Instruct leads
Qwen2.5 72B Instruct: 41.0 (#213), Qwen2.5-Coder-32B: 37.8 (#235)
| Benchmark | Qwen2.5 72B Instruct | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Non-English | 1252 | 1205 |
| LMArena Chinese | 1272 | 1222 |
| LMArena Russian | 1264 | 1228 |
| LMArena French | 1280 | — |
| LMArena German | 1234 | — |
| LMArena Japanese | 1180 | — |
| LMArena Korean | 1188 | — |
| LMArena Spanish | 1256 | — |
Instruction Following Qwen2.5 72B Instruct leads
Qwen2.5 72B Instruct: 65.5 (#221), Qwen2.5-Coder-32B: 61.4 (#245)
| Benchmark | Qwen2.5 72B Instruct | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Instruction Following | 1254 | 1223 |
| LiveBench Instruction Following | — | 58.7% |
| IFEval | 80.6% | — |
Long Context Too close to call
Qwen2.5 72B Instruct: 38.9 (#188), Qwen2.5-Coder-32B: 38.0 (#208)
| Benchmark | Qwen2.5 72B Instruct | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Longer Query | 1282 | 1251 |
Writing & Preference Qwen2.5 72B Instruct leads
Qwen2.5 72B Instruct: 46.7 (#215), Qwen2.5-Coder-32B: 41.6 (#240)
| Benchmark | Qwen2.5 72B Instruct | Qwen2.5-Coder-32B |
|---|---|---|
| LMArena Text | 1269 | 1230 |
| LMArena Creative Writing | 1221 | 1174 |
| LMArena Multi-Turn | 1272 | 1222 |
| WildBench | 80.2% | — |
| LiveBench Language | — | 23.3% |
Frequently asked questions
Is Qwen2.5 72B Instruct better than Qwen2.5-Coder-32B?
Qwen2.5-Coder-32B is the stronger model overall, scoring 33.4 to 31.9 on the Noometry Index.
Which is cheaper, Qwen2.5 72B Instruct or Qwen2.5-Coder-32B?
Qwen2.5-Coder-32B is cheaper. It lists at $0.66 per million input tokens and $1 per million output tokens; Qwen2.5 72B Instruct lists at $1.40 and $5.60.
Is Qwen2.5 72B Instruct or Qwen2.5-Coder-32B better for coding?
Qwen2.5 72B Instruct scores higher on coding benchmarks: 33.2 versus 22.6 in the Noometry coding category.
Which has the bigger context window?
Qwen2.5 72B Instruct does, with 131K tokens against 33K.
How many benchmarks do Qwen2.5 72B Instruct and Qwen2.5-Coder-32B share?
19 benchmarks have published results for both models. Qwen2.5 72B Instruct has 43 scored results on Noometry and Qwen2.5-Coder-32B has 31.