Cloudflare has expanded its Clef decision model family with Clef-omni, an open-weight model that natively handles audio, video, images and text in one pipeline. The company announced the addition alongside a price cut for Clef-flash, its faster sibling, bringing it below a competing model called Jev. Techmeme flagged the news as a Cloudflare product announcement.
The move puts Cloudflare deeper into the layered market for AI models that do more than chat. A single pipeline for four input types matters for developers building on Cloudflare's edge, because most rivals still route audio or video through separate models stitched together.
Clef-omni Specs and What Open-Weight Means
Open-weight means the trained parameters are published, so companies can run the model on their own hardware or inside Cloudflare's network rather than relying only on a hosted API. That is a different bet from closed frontier labs, and it matters for teams with data-residency or cost constraints.
Cloudflare describes Clef-omni as natively processing all four input types in a single pipeline, rather than converting audio and video to text first. That design choice usually cuts latency and error stacking, though the company has not published benchmark numbers in the announcement.
- Name: Clef-omni, part of the Clef decision model family
- Inputs: audio, video, images and text, natively in one pipeline
- Licensing: open-weight
- Companion model: Clef-flash, now priced below Jev
