---
format: "aidr-story-markdown/v1"
id: "4067452015e731b45b62c72357e8f676818e90a9565efdb6ed05a1ca60679d8b"
canonical_url: "https://aidr.today/40674520?lang=en"
title: "FrameForge: Harnessing Priors of Video Foundation Models for Unified Image Generation"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-11-02T00:00:00.000Z"
category: "Research"
topics: ["reasoning"]
source_urls: ["https://huggingface.co/papers/2610.36545"]
summary: "FrameForge harnesses priors of video foundation models for unified image generation. It models conditional generation through input-adaptive virtual temporal relations. FrameForge includes Mixture of Temporal Experts and Chain of Temporal Reasoning."
---

# FrameForge: Harnessing Priors of Video Foundation Models for Unified Image Generation

> [Open the canonical story](<https://aidr.today/40674520?lang=en>)

**Published:** 2026-11-02T00:00:00.000Z
**Category:** Research
**Topics:** reasoning

## Summary

FrameForge harnesses priors of video foundation models for unified image generation\. It models conditional generation through input\-adaptive virtual temporal relations\. FrameForge includes Mixture of Temporal Experts and Chain of Temporal Reasoning\.

## Sources

- [Story source](<https://huggingface.co/papers/2610.36545>)