Cloudflare releases Clef-omni, a 30B multimodal decision model for text, image, audio and video
Cloudflare launched Clef-omni on October 9, 2026, a new open-weight multimodal decision model that processes text, images, audio and video in a single call. Built on the Qwen3-Omni-30B-A3B-Instruct backbone, Clef-omni uses a mixture-of-experts (MoE) architecture with 3 billion active parameters and a 64,000-token context window on Workers AI. It supports up to four images, four audio clips, and…
Key points
- Clef-omni processes text, images, audio and video in one decision call, using a 30B MoE model with 64k context
- Cloudflare claims 97.7 macro-F1 on CLINC150+OOS but lags behind Clef on reasoning and workflow benchmarks
- Hosted Clef is now 1.7x–2.0x faster, Clef-flash costs $0.038, and Clef-omni is open-weight under Apache-2.0
Cloudflare claims Clef-omni excels in intent classification, achieving a 97.7 macro-F1 score on CLINC150+OOS, but trails its own text-focused Clef model on reasoning and workflow tasks. The model’s pricing starts at $0.15 per million tokens, while Clef-flash pricing dropped to $0.038, now under TypeSafe’s Jev model. Cloudflare also improved hosted Clef performance by 1.7x to 2.0x using SGLang, with no new weights released.
The story so far
3 episodes →- Cloudflare releases Clef-omni, a 30B multimodal decision model for text, image, audio and videothis story
Meet Cloudflare Clef-omni: A 30B-A3B Multimodal Decision Model for Audio, Video, Image and Text
MarkTechPost · 11 October 2026
Loading the full article…
This text was published by MarkTechPost and written by Asif Razzaq. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Generative AI & Models
All →- Microsoft releases Decision-1 AI model starting at $0.042 per million input tokens · 1 src
- Google launches Gemini 4 Argon model · 14 src
- Z.ai’s GLM-Legal scores 19.2% on Harvey LAB, outpacing Claude Fable 5 and Kimi K3 · 1 src
- vLLM Semantic Router releases Decision 3.0 multimodal models with 5 sizes under Apache-2.0 · 1 src
- Meta introduces Muse Wonder and Muse Wonder Full models · 15 src
Comments
via GitHub Discussions