Model comparison
Qwen2.5 72B Instruct vs Qwen3.7 Max
Qwen3.7 Max is the stronger model overall, scoring 51.4 to 31.5 on the Noometry Index. Qwen2.5 72B Instruct costs 1.5× less per token, which makes it the better buy when Qwen3.7 Max's lead doesn't matter for your workload.
Last verified . 17 shared benchmarks.
Alibaba (Qwen)Summary
- They share 17 benchmarks with published results for both. Qwen2.5 72B Instruct scores higher in 1 category and Qwen3.7 Max in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3.7 Max leads 62.3 to 18.1.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 8.1% for Qwen2.5 72B Instruct and 95.6% for Qwen3.7 Max.
- Qwen2.5 72B Instruct is cheaper at $1.40 / $5.60 per million input/output tokens, against $2.50 / $7.50 for Qwen3.7 Max.
- Qwen3.7 Max accepts more context: 1M tokens versus 131K.
- Qwen2.5 72B Instruct has downloadable open weights; the other is API-only.
Side by side
| Qwen2.5 72B Instruct | Qwen3.7 Max | |
|---|---|---|
| Provider | Alibaba (Qwen) | Alibaba (Qwen) |
| Noometry Index | 31.5 | 51.4 |
| Released | 2024-09 | 2026-05-19 |
| Weights | Open | Proprietary |
| Context window | 131K | 1M |
| Max output | 8K | 131K |
| Input $ / M tokens | $1.40 | $2.50 |
| Output $ / M tokens | $5.60 | $7.50 |
| Results tracked | 43 | 33 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.7 Max leads
Qwen2.5 72B Instruct: 32.8 (#256), Qwen3.7 Max: 50.2 (#43)
| Benchmark | Qwen2.5 72B Instruct | Qwen3.7 Max |
|---|---|---|
| LMArena Coding | 1293 | 1498 |
| SWE-bench Verified | — | 77.3% |
| LMArena WebDev | — | 1515 |
| SciCode | — | 48.8% |
| WeirdML | 16% | — |
| BigCodeBench Instruct | 45.8% | — |
| BigCodeBench Complete | 55.9% | — |
| ALE-Bench | — | 1,189 |
Agentic & Tool Use Too close to call
Qwen2.5 72B Instruct: 22.0 (#124), Qwen3.7 Max: 21.8 (#129)
| Benchmark | Qwen2.5 72B Instruct | Qwen3.7 Max |
|---|---|---|
| TheAgentCompany | 5.7% | — |
| BALROG | 16.2% | — |
| GBAEval | — | 0.4% |
| METR Time Horizons | 35.8% | — |
Reasoning Qwen3.7 Max leads
Qwen2.5 72B Instruct: 21.4 (#197), Qwen3.7 Max: 48.4 (#36)
| Benchmark | Qwen2.5 72B Instruct | Qwen3.7 Max |
|---|---|---|
| LMArena Hard Prompts | 1271 | 1483 |
| DTBench | 62.9% | 92.3% |
| LMCA | 13.4% | 44% |
| Epoch Capabilities Index | 129.01 | 153.74 |
| SimpleBench | — | 70.4% |
| NYT Connections (extended) | — | 85.1% |
| CritPt | — | 13.4% |
| Chess Puzzles | — | 19% |
| EBR-Bench | — | 9.5% |
| Mystery Game Puzzles | — | 32% |
| BIG-Bench Hard | 79.8% | — |
| ForecastBench | 57.5 | — |
| HellaSwag | 84.8% | — |
| PIQA | 82.6% | — |
| WinoGrande | 82.3% | — |
Math Qwen3.7 Max leads
Qwen2.5 72B Instruct: 18.1 (#281), Qwen3.7 Max: 62.3 (#26)
| Benchmark | Qwen2.5 72B Instruct | Qwen3.7 Max |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 8.1% | 95.6% |
| LMArena Math | 1283 | 1488 |
| FrontierMath (Tiers 1-3) | — | 64.6% |
| FrontierMath Tier 4 | — | 34.1% |
| ProofBench | — | 26% |
| Omni-MATH | 33% | — |
| MATH Level 5 | 63.2% | — |
Knowledge Qwen3.7 Max leads
Qwen2.5 72B Instruct: 26.9 (#247), Qwen3.7 Max: 61.7 (#24)
| Benchmark | Qwen2.5 72B Instruct | Qwen3.7 Max |
|---|---|---|
| GPQA Diamond | 49.1% | 90.9% |
| LMArena Expert | 1245 | 1488 |
| SimpleQA Verified | — | 55.8% |
| MMLU-Pro | 63.1% | — |
| Confabulations | 19.1% | — |
| GPQA (HELM) | 42.6% | — |
| ARC (AI2) Challenge | 94.5% | — |
| MMLU | 85.3% | — |
| TriviaQA | 71.9% | — |
Multilingual Qwen3.7 Max leads
Qwen2.5 72B Instruct: 41.2 (#208), Qwen3.7 Max: 57.1 (#13)
| Benchmark | Qwen2.5 72B Instruct | Qwen3.7 Max |
|---|---|---|
| LMArena Non-English | 1252 | 1475 |
| LMArena Chinese | 1272 | 1533 |
| LMArena Russian | 1264 | 1482 |
| LMArena French | 1279 | — |
| LMArena German | 1234 | — |
| LMArena Japanese | 1180 | — |
| LMArena Korean | 1188 | — |
| LMArena Spanish | 1255 | — |
Instruction Following Qwen3.7 Max leads
Qwen2.5 72B Instruct: 65.7 (#216), Qwen3.7 Max: 76.9 (#34)
| Benchmark | Qwen2.5 72B Instruct | Qwen3.7 Max |
|---|---|---|
| LMArena Instruction Following | 1254 | 1460 |
| IFEval | 80.6% | — |
Long Context Qwen3.7 Max leads
Qwen2.5 72B Instruct: 38.9 (#183), Qwen3.7 Max: 45.4 (#34)
| Benchmark | Qwen2.5 72B Instruct | Qwen3.7 Max |
|---|---|---|
| LMArena Longer Query | 1282 | 1484 |
Writing & Preference Qwen3.7 Max leads
Qwen2.5 72B Instruct: 46.9 (#210), Qwen3.7 Max: 65.2 (#52)
| Benchmark | Qwen2.5 72B Instruct | Qwen3.7 Max |
|---|---|---|
| LMArena Text | 1269 | 1476 |
| LMArena Creative Writing | 1221 | 1450 |
| LMArena Multi-Turn | 1272 | 1482 |
| WildBench | 80.2% | — |
| EQ-Bench 4 | — | 1110 |
Frequently asked questions
Is Qwen2.5 72B Instruct better than Qwen3.7 Max?
Qwen3.7 Max is the stronger model overall, scoring 51.4 to 31.5 on the Noometry Index. Qwen2.5 72B Instruct costs 1.5× less per token, which makes it the better buy when Qwen3.7 Max's lead doesn't matter for your workload.
Which is cheaper, Qwen2.5 72B Instruct or Qwen3.7 Max?
Qwen2.5 72B Instruct is cheaper. It lists at $1.40 per million input tokens and $5.60 per million output tokens; Qwen3.7 Max lists at $2.50 and $7.50.
Is Qwen2.5 72B Instruct or Qwen3.7 Max better for coding?
Qwen3.7 Max scores higher on coding benchmarks: 50.2 versus 32.8 in the Noometry coding category.
Which has the bigger context window?
Qwen3.7 Max does, with 1M tokens against 131K.
How many benchmarks do Qwen2.5 72B Instruct and Qwen3.7 Max share?
17 benchmarks have published results for both models. Qwen2.5 72B Instruct has 43 scored results on Noometry and Qwen3.7 Max has 33.