Best cell per row highlighted. Null means undisclosed — never counted as zero.
| Provider | Alibaba Cloud | OpenAI |
| Family / tier | Qwen2.5-Coder · Coder | GPT-5 · Medium |
| Status | GA | GA |
| Released | 2024-11-12 | 2026-03-05 |
| Knowledge cutoff | 2024-08 | 2025-08 |
| Context window | 131K | 1.1M |
| Max output tokens | 8K | 128K |
| Input modalities | text | text, image |
| Reasoning mode | none | optional |
| Open weights | Yes | No |
| Input | $0.08 | $2.5 |
| Output | $0.24 | $15 |
| Cached input | — | $0.25 |
| Batch input | — | $1.25 |
| Free tier | Yes | No |
| Output speed (tok/s) | — | 86.3 |
| Time to first token (s) | — | 1.2 |
| Latency tier | fast | medium |
| Trains on inputs | No | No |
| License | Apache-2.0 | Proprietary |
| Clouds | GCP | Azure OpenAI, Azure AI Foundry |
| Compliance | — | SOC2, GDPR, CCPA, HIPAA |
| Unified score | 8.4 | 8.7 |
| Decision Maker | 8.5 | 9 |
| Domain Strategist | 8 | 8.7 |
| Finance Lead | 9 | 9 |
| Domain Practitioner | 9.5 | 9 |
| Power User | 8.5 | 8.5 |
| Skeptic | 7.5 | 8 |
| Value score | 9.5 | 8.5 |