Skip to main content

Open models, fine-tuned, against a frontier API.

Narrow, high-volume tasks, across zero-shot, few-shot and fine-tuned conditions.

Every model and method, scored

Model Method Metric ↓
Open-weight, fine-tuned (Apache 2.0)
Hosted API, cost tier

Where the errors go

Reproduce any number

These are public-data results. See if it holds for your task.

Book a call →