Which model is best? The honest answer is that it depends on the question. So instead of guessing, we let you see for yourself.
One prompt, many models
In the Arena, you pick two to four models, type one prompt, and every model answers at the same time in its own column. You watch them stream in side by side and judge the results with your own eyes.
Why it is useful
- You stop wondering which model to trust for a given task and simply look.
- You can pit a fast local model against a large cloud model and see if the upgrade is worth it.
- It is a great way to learn the strengths of each model without reading a benchmark chart.
The Arena is a Pro feature, and it turned out to be the launch pad for two more ideas: combining the answers, and scoring them. Those became SDL Consensus and AI Score.