Output Speed
Median output tokens generated per second.
Output Speed is an operational performance metric, not an intelligence benchmark. Artificial Analysis defines it as the average number of output tokens received per second after the first token arrives, using OpenAI tokens as the standard unit.
Test type: Hosted endpoint performance measurement. Higher values mean faster streaming after response start.
Coverage
300 models have this metric.
758.7 tok/s
Current leader: Mercury 2
Project links
Values come from Artificial Analysis performance data in the committed snapshot.
Top Speed Models
Top models ranked by Speed.
Leaderboard
| Rank | Model | Creator | Value | Speed | Blended Price |
|---|---|---|---|---|---|
| #1 | Mercury 2 | Inception | 758.7 tok/s | 758.7 tok/s | $0.375/M |
| #2 |
