Cloudflare's Clef-omni handles audio, video, and images in one pipeline with 2x faster inference
Cloudflare expands its decision model family with a new multimodal AI, alongside faster inference and reduced pricing for existing models.
Key points
- Clef-omni natively processes audio, video, images, and text in a single pipeline.
- Clef inference speeds have increased by up to 2.0x.
- Pricing for Clef-flash has been lowered.
Cloudflare has released Clef-omni, a new addition to its Clef decision model family that supports full multimodality. According to first-hand research published by Cloudflare, this model natively processes audio, video, images, and text within a single pipeline. The company also announced performance improvements and cost reductions for its existing Clef products.
Root cause
The development focuses on expanding the capabilities of the Clef decision model family. Cloudflare engineers built Clef-omni to handle multiple input types simultaneously, removing the need for separate processing steps for different media formats. This architectural change allows for more efficient handling of complex data streams.
Attack path
This release does not involve a security vulnerability or an attack path. It is a product launch detailing new features and performance enhancements. The changes aim to improve the utility and efficiency of the decision models for developers and enterprises using Cloudflare’s AI services.
Affected versions
The announcement applies to the Clef decision model family. Specifically, it introduces Clef-omni as a new model. It also updates the performance characteristics of the existing Clef model and the pricing structure for Clef-flash. No specific version numbers are provided in the source material.
Mitigation
No mitigation is required as this is a product feature announcement, not a security incident. Users may choose to adopt the new Clef-omni model or update their workflows to take advantage of the faster Clef inference speeds and lower Clef-flash costs.
What to do and how to stay safe: Cloudflare
- Review current AI model usage to determine if multimodal processing would benefit your applications.
- Test the new Clef-omni model in a staging environment to evaluate performance gains.
- Monitor the updated pricing for Clef-flash to adjust budget forecasts for AI inference.
- Verify that your integration code supports the new multimodal input formats before deployment.
Step-by-step guide: Phishing vs Vishing: How Audio Bypasses Your Defences
General security guidance from the TechSec Observer newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
What types of data can Clef-omni process?
Clef-omni natively processes audio, video, images, and text in a single pipeline, according to Cloudflare.
How much faster is the Clef model now?
Cloudflare states that Clef inference speeds have been boosted by up to 2.0x.
Are there any price changes for Clef models?
Yes, Cloudflare has lowered the pricing for Clef-flash, as announced in their blog post.
