Token cost
USD per 1M tokens, 50% input + 50% output. Lower is better.
0 models in this catalog are created by Microsoft Azure; 9 additional models are available through its hosting offerings. The newest dated release listed is GPT-6.1 Sol (Sep 29, 2026).
USD per 1M tokens, 50% input + 50% output. Lower is better.
Tokens per second. Higher is better.
Seconds until the first token. Lower is better.
| Model | Date released | Model details |
|---|---|---|
| GPT-6.1 Sol | View model | |
| Claude Sonnet 5.5 | View model | |
| Claude Opus 5.5 | View model | |
| GPT-6 Astra | View model | |
| GPT-5.6 Sol | View model | |
| Claude Sonnet 5 | View model | |
| Claude Fable 5 | View model | |
| Claude Opus 4.7 | View model | |
| Claude Opus 4.6 | View model |