What is the Cloudflare multimodal decision model Clef-omni?
Key points
- Processes text, images, audio, and video simultaneously
- Priced at $0.15 per million tokens on Workers AI
- Scores a 21-second video in approximately 1.5 seconds
- Open-weight model under Apache-2.0 license
Sources
FAQ
What inputs does Clef-omni support?
It supports up to 4 images, 4 audio clips (up to 300 seconds each), and 2 video clips (up to 60 seconds each).
How much does Clef-omni cost?
The hosted version on Workers AI costs $0.15 per million tokens.
Video transcript (Hinglish)
Cloudflare ne apna naya multimodal decision model Clef-omni launch kar diya hai. Isse aapki apps kaise badal jayengi, chalo batata hoon lekin pehle AI Jockey ko follow kar lena. Clef-omni ek hi call mein text, images, audio aur video ko score kar sakta hai. Ye model text generate nahi karta balki aapke options ko probability scores deta hai. Developers ke liye fayda ye hai ki ab transcription aur captioning ki zarurat nahi padegi. Isme aap chaar images, chaar audio clips aur do videos tak bhej sakte hain. Speed ki baat karein toh ek ekkis second ka video sirf dedh second mein score ho jata hai. Ye Workers AI par available hai aur iska price sirf zero point one five dollars per million tokens hai. Lekin ek catch hai, ye model reasoning aur workflow tasks mein thoda peeche reh jata hai. Ye model un founders aur developers ke liye best hai jo intent classification par kaam kar rahe hain. Kya aap apne project mein multimodal decision model use karenge? Comment karo DECISION, main bhej deta hoon.