Best cell per row highlighted. Null means undisclosed — never counted as zero.
| Provider | Alibaba Cloud | Alibaba Cloud |
| Family / tier | Qwen2.5 · Large | Qwen2.5 · Large |
| Status | GA | GA |
| Released | 2024-09-18 | 2024-09-18 |
| Knowledge cutoff | 2024-06 | 2024-06 |
| Context window | 131K | 131K |
| Max output tokens | 8K | 8K |
| Input modalities | text | text |
| Reasoning mode | none | none |
| Open weights | Yes | Yes |
| Input | $0.12 | $0.1 |
| Output | $0.3 | $0.25 |
| Cached input | — | — |
| Batch input | — | — |
| Free tier | Yes | Yes |
| Output speed (tok/s) | — | — |
| Time to first token (s) | — | — |
| Latency tier | fast | fast |
| Trains on inputs | No | No |
| License | Qwen | Apache-2.0 |
| Clouds | GCP | GCP |
| Compliance | — | — |
| Unified score | 7.5 | 7.2 |
| Decision Maker | 7.5 | 7 |
| Domain Strategist | 7 | 7 |
| Finance Lead | 9 | 9 |
| Domain Practitioner | 8 | 8 |
| Power User | 7 | 6.5 |
| Skeptic | 7.5 | 7.5 |
| Value score | 8.5 | 8.5 |
| MMLU | 86.1 | 83.3 |
| MMLU-Pro | 71.1 | 69 |
| MATH-500 | 83.1 | 83.1 |
| HumanEval | 86.6 | 88.4 |
Scores link to their sources. Missing cells mean the vendor hasn't disclosed a result — honesty over padding.