Best cell per row highlighted. Null means undisclosed — never counted as zero.
| Provider | OpenAI | OpenAI | OpenAI |
| Family / tier | GPT-5 · Medium | GPT-5 · Pro | GPT-5 · Pro |
| Status | GA | GA | GA |
| Released | 2026-03-04 | 2026-04-23 | 2026-04-22 |
| Knowledge cutoff | 2025-08 | 2025-12 | 2025-12 |
| Context window | 1.1M | 1.1M | 1.1M |
| Max output tokens | 128K | 128K | 128K |
| Input modalities | text, image | text, image | text, image |
| Reasoning mode | optional | always | optional |
| Open weights | No | No | No |
| Input | $2.5 | $30 | $5 |
| Output | $15 | $180 | $30 |
| Cached input | $0.25 | — | $0.5 |
| Batch input | $1.25 | $15 | $2.5 |
| Free tier | No | No | No |
| Output speed (tok/s) | 86.3 | — | 68.7 |
| Time to first token (s) | 1.2 | — | 64 |
| Latency tier | medium | slow | slow |
| Trains on inputs | No | No | No |
| License | Proprietary | Proprietary | Proprietary |
| Clouds | Azure OpenAI, Azure AI Foundry | Azure OpenAI, Azure AI Foundry | Azure OpenAI, Azure AI Foundry |
| Compliance | SOC2, GDPR, CCPA, HIPAA | SOC2, GDPR, CCPA, HIPAA | SOC2, GDPR, CCPA, HIPAA |
| Unified score | 8.7 | 7.7 | 8.7 |
| Decision Maker | 9 | 8 | 9 |
| Domain Strategist | 8.7 | 7.5 | 8.8 |
| Finance Lead | 9 | 6 | 7 |
| Domain Practitioner | 9 | 7 | 9.2 |
| Power User | 8.5 | 8 | 9 |
| Skeptic | 8 | 7 | 7.8 |
| Value score | 8.5 | 5 | 6.5 |
| MMLU-Pro | — | — | 92.4 |
| GPQA Diamond | 92.8 | 94 | 93.6 |
| SWE-bench Verified | — | 89 | 88.7 |
| MMMU | 81.2 | — | 81.2 |
| TAU-bench | — | — | 98 |
| Terminal-Bench | 75.1 | — | 82.7 |
| MRCR Long Context | 36.6 | — | 74 |
| Humanity's Last Exam | 39.8 | — | 41.4 |
| LMArena Elo | 1484 | — | 1476 |
| Artificial Analysis Index | 57 | — | 60 |
Scores link to their sources. Missing cells mean the vendor hasn't disclosed a result — honesty over padding.