Why this directory exists
A model announcement is only the beginning. Developers need to know where the documentation lives, what an API accepts, what it costs and what has not yet been established. We organize those details into model briefings and practical guides.
Jev is our first research cluster. This directory is designed to cover models from many providers, including text, image, video, coding, reasoning and agent models.
How we verify information
- Start with the provider’s website, documentation, release notes and repositories.
- Record a source and review date for consequential facts.
- Distinguish official documentation, provider claims and our own testing.
- Leave missing information visibly unconfirmed instead of filling gaps with guesses.
- Publish new guides only when they answer a concrete question with useful evidence.
What a verification label means
“Documented” means a claim is supported by the linked primary source at the stated review date. It does not mean that we have tested the service, audited the provider, or guaranteed ongoing availability. An unconfirmed status means our reviewed sources did not establish the fact; it is not a negative claim.
Examples and benchmarks
Code is labeled with its runtime and testing status. We do not publish fabricated response values or benchmark numbers. Future benchmark articles must disclose the dataset, prompts, model versions, date, costs and evaluation method.
Independence and corrections
We are not affiliated with TypeSafe AI or other model providers. This edition contains no sponsored placements or affiliate links. Review dates are visible on the model and guide pages. When official information changes, corrections should update the affected facts, citations and review dates together.
Contact
AI Model Brief is operated by jessie. For editorial questions, corrections or privacy requests, contact jessiejessie969@gmail.com.