Measured over the last 30 days
- Models Tracked
- 33
- Avg Tokens / Second
- 23.04
- Avg Time to First Token (ms)
- 8598.18
- Last Updated
- Sep 23, 2026
All openai models
33 live · one row per model · fastest first · 5 retired lanes not listed| Model | Avg Toks/Sec | Min | Max | Avg TTF (ms) |
|---|---|---|---|---|
| GPT-6 Luna Pro | 53.70 | 53.70 | 53.70 | 31000.00 |
| gpt-4.1 | 41.90 | 30.70 | 52.00 | 790.00 |
| GPT-6 Sol Pro | 41.20 | 41.20 | 41.20 | 10700.00 |
| GPT-5.4-nano | 40.40 | 32.30 | 53.30 | 1640.00 |
| o1 | 39.30 | 5.77 | 73.10 | 3100.00 |
| GPT-5.4-mini | 34.90 | 26.30 | 48.80 | 1470.00 |
| GPT-5.1 | 33.80 | 16.20 | 50.20 | 1540.00 |
| o3 | 32.00 | 18.50 | 48.50 | 2210.00 |
| gpt-4.1-nano | 31.40 | 19.50 | 37.50 | 1420.00 |
| gpt-4o-mini | 31.10 | 26.30 | 41.20 | 1540.00 |
| GPT-5.5 | 25.90 | 20.20 | 31.00 | 1780.00 |
| gpt-4.1-mini | 25.70 | 19.50 | 34.30 | 1570.00 |
| gpt-4o | 24.10 | 9.63 | 34.40 | 1870.00 |
| GPT-5.2 | 22.90 | 19.60 | 25.30 | 2080.00 |
| GPT-5.6-luna | 21.90 | 11.60 | 31.40 | 2650.00 |
| GPT-5.4 | 21.50 | 19.80 | 23.60 | 1970.00 |
| GPT Luna Latest | 20.90 | 19.70 | 22.10 | 2810.00 |
| GPT-5.3-codex | 20.10 | 14.20 | 28.40 | 1880.00 |
| gpt-3.5-turbo | 19.10 | 13.20 | 23.00 | 2960.00 |
| GPT-5.6-terra | 18.10 | 10.60 | 26.50 | 2150.00 |
| GPT-5.6-sol | 17.60 | 11.20 | 26.60 | 2390.00 |
| gpt-4 | 16.80 | 10.50 | 20.20 | 2780.00 |
| GPT Terra Latest | 15.90 | 13.90 | 18.00 | 3430.00 |
| o3-pro | 15.80 | 9.63 | 20.00 | 26300.00 |
| GPT-6 Sol | 15.40 | 15.40 | 15.40 | 3610.00 |
| gpt-4-turbo | 14.70 | 12.10 | 16.60 | 2920.00 |
| GPT Sol Latest | 13.10 | 11.80 | 14.50 | 4210.00 |
| GPT-5.2-pro | 12.90 | 10.80 | 15.20 | 33100.00 |
| GPT-6 Luna | 10.70 | 10.70 | 10.70 | 5580.00 |
| GPT-5.5-pro | 9.79 | 3.04 | 16.80 | 31700.00 |
| GPT-6 Astra Pro | 7.69 | 3.65 | 11.70 | 7970.00 |
| GPT Astra Latest | 6.15 | 6.15 | 6.15 | 5120.00 |
| GPT-5.4-pro | 3.96 | 2.77 | 5.55 | 77500.00 |
Frequently Asked Questions
Which openai model is fastest?
Based on recent tests, GPT-6 Luna Pro shows the highest average throughput among tracked openai models.
How many recent measurements feed this dashboard?
This provider summary aggregates 199 individual prompts measured across 152 monitoring runs over the past month.