Today for AI
HOT RADAR
llmHEAT 4.2°

AI CLUSTERED EVENT · 10/10/2026

Cloudflare Open-Sources Clef-omni: A Qwen3-Based Decision Model Handling Audio and Video in Single API Calls with 130ms Latency

1 reports archived1 independent sourcesupdated 10/10/2026, 14:26:25
Synthesis & Latest Updates
1 Sources Cross-Validated

Cloudflare has released Clef-omni, an open-weight multimodal decision model capable of processing text, image, audio, and video inputs via a single API call. Built on Qwen3-Omni-30B-A3B-Instruct, it eliminates the need for separate speech-to-text or video frame extraction pipelines for structured decision tasks. Benchmarks indicate a median response time of approximately 130 milliseconds for text requests, with weights now available on Hugging Face.

LATEST/Cloudflare has released Clef-omni, an open-weight multimodal decision model capable of processing text, image, audio, and video inputs via a single API call. Built on Qwen3-Omni-30B-A3B-Instruct, it eliminates the need for separate speech-to-text or video frame extraction pipelines for structured decision tasks. Benchmarks indicate a median response time of approximately 130 milliseconds for text requests, with weights now available on Hugging Face.

HEAT TRENDHourly heat curve

46 fully observed hours
Now
4.2
Peak
11.815:00
24h change
-42%

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

TIMELINECoverage timeline

Total 1 reports · Latest first
  1. IT之家 · 智能时代T2·68 pts
    • Supports wav/mp3/mp4/webm formats, enabling end-to-end single API calls for audio, video, text, and images.
    • Fine-tuned from Qwen3-Omni-30B-A3B-Instruct, focusing on structured decision-making rather than general text generation.
    • High performance: ~130ms median response for text, and ~1.5s to score a 21-second video with audio.