Model comparison
Qwen2.5 72B Instruct vs Qwen3.6 Plus
Qwen3.6 Plus is the stronger model overall, scoring 47.5 to 31.9 on the Noometry Index.
Last verified . 22 shared benchmarks.
Alibaba (Qwen)Summary
- They share 22 benchmarks with published results for both. Qwen2.5 72B Instruct scores higher in 0 categories and Qwen3.6 Plus in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3.6 Plus leads 51.8 to 19.3.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 8.1% for Qwen2.5 72B Instruct and 93.3% for Qwen3.6 Plus.
- Qwen3.6 Plus is cheaper at $0.50 / $3 per million input/output tokens, against $1.40 / $5.60 for Qwen2.5 72B Instruct.
- Qwen3.6 Plus 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 | Qwen3.6 Plus | |
|---|---|---|
| Provider | Alibaba (Qwen) | Alibaba (Qwen) |
| Noometry Index | 31.9 | 47.5 |
| Released | 2024-09 | 2026-03-31 |
| Weights | Open | Proprietary |
| Context window | 131K | 1M |
| Max output | 8K | 66K |
| Input $ / M tokens | $1.40 | $0.50 |
| Output $ / M tokens | $5.60 | $3 |
| Results tracked | 43 | 37 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Qwen3.6 Plus leads
Qwen2.5 72B Instruct: 33.2 (#260), Qwen3.6 Plus: 40.8 (#130)
| Benchmark | Qwen2.5 72B Instruct | Qwen3.6 Plus |
|---|---|---|
| LMArena Coding | 1292 | 1467 |
| SWE-bench Verified | — | 57.9% |
| LMArena WebDev | — | 1461 |
| SciCode | — | 40.7% |
| WeirdML | 16% | — |
| BigCodeBench Instruct | 45.8% | — |
| BigCodeBench Complete | 55.9% | — |
| ALE-Bench | — | 670.15 |
Agentic & Tool Use Not comparable
Qwen2.5 72B Instruct: 22.1 (#133), Qwen3.6 Plus: —
| Benchmark | Qwen2.5 72B Instruct | Qwen3.6 Plus |
|---|---|---|
| TheAgentCompany | 5.7% | — |
| BALROG | 16.2% | — |
| METR Time Horizons | 35.8% | — |
| Vending-Bench 2 | — | 5,115 |
Reasoning Qwen3.6 Plus leads
Qwen2.5 72B Instruct: 22.3 (#199), Qwen3.6 Plus: 29.3 (#93)
| Benchmark | Qwen2.5 72B Instruct | Qwen3.6 Plus |
|---|---|---|
| LMArena Hard Prompts | 1271 | 1449 |
| DTBench | 62.9% | 81.9% |
| LMCA | 13.4% | 33.1% |
| Epoch Capabilities Index | 129 | 147.65 |
| NYT Connections (extended) | — | 60.3% |
| CritPt | — | 2.9% |
| Chess Puzzles | — | 17% |
| Thematic Generalization | — | 59.5% |
| Mystery Game Puzzles | — | 12% |
| BIG-Bench Hard | 79.8% | — |
| ForecastBench | 57.5 | — |
| HellaSwag | 84.8% | — |
| PIQA | 82.6% | — |
| WinoGrande | 82.3% | — |
Math Qwen3.6 Plus leads
Qwen2.5 72B Instruct: 19.3 (#287), Qwen3.6 Plus: 51.8 (#54)
| Benchmark | Qwen2.5 72B Instruct | Qwen3.6 Plus |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 8.1% | 93.3% |
| LMArena Math | 1283 | 1450 |
| FrontierMath (Tiers 1-3) | — | 38.2% |
| Omni-MATH | 33% | — |
| MATH Level 5 | 63.2% | — |
| FrontierMath (Feb 2025 set) | — | 26.2% |
| FrontierMath Tier 4 (v1) | — | 8.3% |
Knowledge Qwen3.6 Plus leads
Qwen2.5 72B Instruct: 27.0 (#253), Qwen3.6 Plus: 56.1 (#45)
| Benchmark | Qwen2.5 72B Instruct | Qwen3.6 Plus |
|---|---|---|
| GPQA Diamond | 49.1% | 88.4% |
| LMArena Expert | 1245 | 1454 |
| SimpleQA Verified | — | 44.1% |
| MMLU-Pro | 63.1% | — |
| Confabulations | 19.1% | — |
| GPQA (HELM) | 42.6% | — |
| ARC (AI2) Challenge | 94.5% | — |
| MMLU | 85.3% | — |
| TriviaQA | 71.9% | — |
Multilingual Qwen3.6 Plus leads
Qwen2.5 72B Instruct: 41.0 (#213), Qwen3.6 Plus: 53.3 (#70)
| Benchmark | Qwen2.5 72B Instruct | Qwen3.6 Plus |
|---|---|---|
| LMArena Non-English | 1252 | 1424 |
| LMArena Chinese | 1272 | 1477 |
| LMArena French | 1280 | 1455 |
| LMArena German | 1234 | 1452 |
| LMArena Japanese | 1180 | 1389 |
| LMArena Korean | 1188 | 1379 |
| LMArena Russian | 1264 | 1434 |
| LMArena Spanish | 1256 | 1432 |
Instruction Following Qwen3.6 Plus leads
Qwen2.5 72B Instruct: 65.5 (#221), Qwen3.6 Plus: 75.0 (#74)
| Benchmark | Qwen2.5 72B Instruct | Qwen3.6 Plus |
|---|---|---|
| LMArena Instruction Following | 1254 | 1425 |
| IFEval | 80.6% | — |
Long Context Qwen3.6 Plus leads
Qwen2.5 72B Instruct: 38.9 (#188), Qwen3.6 Plus: 45.2 (#49)
| Benchmark | Qwen2.5 72B Instruct | Qwen3.6 Plus |
|---|---|---|
| LMArena Longer Query | 1282 | 1439 |
| CL-bench | — | 20.3% |
Writing & Preference Qwen3.6 Plus leads
Qwen2.5 72B Instruct: 46.7 (#215), Qwen3.6 Plus: 62.2 (#82)
| Benchmark | Qwen2.5 72B Instruct | Qwen3.6 Plus |
|---|---|---|
| LMArena Text | 1269 | 1437 |
| LMArena Creative Writing | 1221 | 1404 |
| LMArena Multi-Turn | 1272 | 1438 |
| WildBench | 80.2% | — |
Frequently asked questions
Is Qwen2.5 72B Instruct better than Qwen3.6 Plus?
Qwen3.6 Plus is the stronger model overall, scoring 47.5 to 31.9 on the Noometry Index.
Which is cheaper, Qwen2.5 72B Instruct or Qwen3.6 Plus?
Qwen3.6 Plus is cheaper. It lists at $0.50 per million input tokens and $3 per million output tokens; Qwen2.5 72B Instruct lists at $1.40 and $5.60.
Is Qwen2.5 72B Instruct or Qwen3.6 Plus better for coding?
Qwen3.6 Plus scores higher on coding benchmarks: 40.8 versus 33.2 in the Noometry coding category.
Which has the bigger context window?
Qwen3.6 Plus does, with 1M tokens against 131K.
How many benchmarks do Qwen2.5 72B Instruct and Qwen3.6 Plus share?
22 benchmarks have published results for both models. Qwen2.5 72B Instruct has 43 scored results on Noometry and Qwen3.6 Plus has 37.