Pith. sign in

REVIEW 4 major objections 5 minor 50 references

Explicit world-state pre-materialization makes long-form film generation parallel and coherent

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-08-01 13:36 UTC pith:WR4RRRXL

load-bearing objection A genuinely new idea in long-form generation—explicit world-state bookkeeping with parallel rendering—but the 5-point superiority claim is scored by the same model family that generates the films, so the margin is not yet credible. the 4 major comments →

arxiv 2607.19038 v1 pith:WR4RRRXL submitted 2026-07-21 cs.CV cs.AI

FilmWorld: Agentic Novel-to-Film Generation through Dynamic Cinematic World Modeling

Jialong Zuo , Haotong Zuo , Shiwei Zhang , Xiang Wang , Chen Li , Nong Sang , Changxin Gao , Xiang Bai This is my paper
classification cs.CV cs.AI
keywords novel-to-film generationdynamic cinematic world modelingmulti-agent video generationworld state externalizationcross-scene consistencynarrative fidelitylatency reductionautomated evaluation
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper argues that the way to generate coherent long films from novels is to model the source as a dynamic cinematic world with persistent, addressable entity states, and to fully resolve how that world changes before any pixels are rendered. It proposes a two-phase agent system that constructs a structured, visually anchored world state from prose, then evolves that state causally through plot events, referencing precomputed state identifiers at every rendering step. The central claimed payoff is that this externalization removes the sequential dependency between shots, cutting the theoretical latency of the rendering-dominant phase from O(N) to O(1) while making cross-scene visual and narrative consistency a matter of state lookup rather than accumulated guesswork. The paper also contributes a difficulty-graded benchmark with nine automatic metrics and reports that its system outperforms five existing agentic video generation baselines, especially on narrative fidelity and cross-scene consistency.

Core claim

On the paper's own terms, the discovery is that novel-to-film generation should be treated as dynamic cinematic world modeling, decomposed into construction (grounding prose into stateful entities) and evolution (causally updating those states across scenes), and that pre-materializing the entire world-state trajectory as a symbolic blueprint before rendering transforms long-form coherence from a sequential, memory-bound problem into a parallel, lookup-bound one. Concretely, every entity (character, location, prop) is assigned a deterministic state identifier at each shot, all visual assets are anchored to those identifiers at first appearance, and the transition of states under plot events

What carries the argument

The load-bearing object is the Dynamic Cinematic World tuple W=(E,X,Φ,T,R): an entity set E (characters, locations, props), a discretized state space X, a state mapping Φ that hashes each concrete state to a unique identifier, a transition function T that updates the world state from plot events, and a rendering function R that conditions each shot on the active state identifiers. The decisive step is the offline materialization of the state trajectory {W_t} via T during construction, so that at evolution time each shot's renderer consumes no other shot's rendered output, which is what yields the O(1)-in-N latency bound.

Load-bearing premise

The paper's measured advantage rests on the assumption that its nine automatic metrics, all implemented by prompting a single multimodal language model that is also used as the system's planner, genuinely track human-judged cinematic quality and fidelity; if that judge systematically favors outputs matching its own stylistic defaults, the reported margins could be inflated.

What would settle it

Run the FilmEval protocol with an independent, instruction-tuned evaluator (for instance a different multimodal model or a blind human panel) on the same 90 generated films; if the system no longer ranks first or the margin over baselines shrinks to near zero on novel fidelity, the central empirical claim is not stable.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • If the state trajectory is correct, cross-shot character and scene coherence become enforced by construction rather than emergent, so errors do not compound with film length.
  • Because rendering is decoupled from sequential state propagation, all keyframes, motion plans, and video segments within a scene can be synthesized in parallel, making the runtime-dominant phase O(1) in the number of shots (or O(N/P) with P parallel workers).
  • Narrative fidelity is improved because plot events are explicitly chained into state transitions, so omissions or hallucinations are detectable and correctable by closed-loop verification.
  • The proposed FilmEval benchmark (15 novels, three difficulty tiers, nine metrics) provides a standardized way to compare long-form generation systems, and the paper reports that its system ranks first on all metric groups and difficulty tiers.
  • Ablations show that removing explicit world-state modeling collapses character consistency, indicating that state externalization is the structural foundation of the observed gains.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The O(1) latency claim is not tied to the specific agentic implementation: any long-form generator that pre-materializes a state trajectory and satisfies the decoupling condition It ⊆ {Wt, L, dt, keyframes} should inherit the same parallel speedup, making the principle broadly applicable.
  • Because the same multimodal language model is used both as the system's planner and as the judge for FilmEval's automatic metrics, the reported margins may partly reflect the judge's preference for outputs that match its own stylistic defaults; an independent, instruction-tuned evaluator would be a stronger test of the empirical claim.
  • The discrete world-state representation (age stage, costume, season, weather) makes consistency easy but quantizes continuous visual evolution such as gradual aging or subtle lighting shifts; blending continuous attributes into the state space is a natural extension that could preserve both consistency and realism.
  • The closed-loop verifier only repairs local deviations from the prescribed state; if a plot event was mis-assigned during construction, the error propagates silently, so upstream parsing correctness remains the dominant risk for faithfulness.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper formalizes novel-to-film generation as dynamic cinematic world modeling, decomposing the task into a Construction Phase that grounds literary prose into a persistent, stateful world and an Evolution Phase that renders shots from a pre-materialized state trajectory. It proposes FilmWorld, a six-agent system implementing this decomposition, and FilmEval, a difficulty-graded benchmark of 15 novels with nine automatic metrics across Cinematic Presentation, Film Consistency, and Novel Fidelity. The main claims are: (i) explicit externalization of world state improves long-range coherence, with FilmWorld outperforming five agentic baselines by 5.19 points overall (Table 1); (ii) state externalization reduces Evolution-Phase latency from O(N) to O(1) (Eqs. 8–10), with measured 5.62× speedup on keyframe generation; and (iii) FilmEval's metrics align with human judgment. The formalization and system design are coherent and the latency argument is internally sound under a bounded-context condition, but the empirical support for the headline margin is weakened by the fact that the automatic judge is the same model family as the generator, and the independent human validation is thin.

Significance. If the empirical claims hold, this is a substantial contribution: it reframes long-form video generation as a world-modeling problem rather than a per-shot synthesis problem, introduces an explicit and addressable state representation that can be shared across agents, and offers a plausible mechanism for parallelizing shot rendering. The paper's strengths include a clean formalization (Section 3), a detailed system design (Section 4), a public benchmark with per-novel scores (Appendix B.1), ablations isolating each component (Table 4), a cross-evaluator stability check (Appendix B.4), a human evaluation study (Section 6.2), a style generalization study (Appendix B.2), and a candid failure analysis (Figure 16, Section E.1). These are genuine assets. The main weakness is that the quantitative superiority claim rests on an automated protocol whose judge shares model ancestry with the system being evaluated; the cross-evaluator re-run mitigates but does not eliminate the concern, and the human validation is too small to independently certify the reported margin. The latency claim is theoretically sound but only partially validated experimentally. These issues are load-bearing for the a

major comments (4)
  1. [§5.3, Table 1, Appendix B.4] The central empirical claim—a 5.19-point overall margin over VideoClaw (Table 1)—is measured entirely by FilmEval's nine automatic metrics, all implemented by prompting Gemini 3.1 Pro (§5.3), the same model family used as FilmWorld's planner and internal verifier. If the judge systematically prefers outputs matching its own planning conventions or stylistic defaults, the margin may not transfer to an independent evaluator. The cross-evaluator re-run in Appendix B.4 (Qwen3.5-Omni, Seed2.0-Lite) shows the system-level ranking is stable (ρ≥0.9), which is reassuring, but all three judges are web-scale MLLMs with overlapping training priors; this does not fully remove the self-preference risk. The only independent validation is the human study (Table 3), which has six annotators, no inter-annotator agreement, and no confidence interval on the decisive FilmWorld-vs-VideoClaw margin. Please rep
  2. [§4.4, Eq. (10)] The State Decoupling Condition is stated as: 'This reduction in complexity holds iff the information set I_t consumed by the renderer at shot t contains no rendered outputs from other shots.' However, Eq. (10) immediately includes {k_j}_{j∈J_t}, where k_j is the keyframe image of shot j, and J_t is a window of other shots. In the actual pipeline, the sliding-window planner consumes keyframes of context shots, so the information set does contain rendered outputs of other shots at the planning stage. The O(1) latency argument is preserved if the keyframe context is bounded (w = g+2c << N), but the text and Eq. (10) are internally inconsistent as written. Please clarify what counts as a 'rendered output' (final video segment vs. keyframe), and state explicitly that bounded keyframe context is permitted because its size does not grow with N.
  3. [§4.4, Figure 10] The experimental efficiency validation measures only the keyframe-generation stage, which is described as the most cleanly parallelizable component. The dynamic-planning and video-rendering stages—which are also part of the claimed O(1) Evolution-Phase latency—are not measured end-to-end. The statement that the remaining stages 'share this shot-independent structure and should exhibit analogous gains' is a prediction, not a measurement. Because API rate limits, memory constraints, and batching behavior can make full-pipeline scaling worse than the theoretical O(N/P), please either measure the full Evolution Phase or explicitly qualify the experimental claim to the keyframe stage while keeping the O(1) bound as a theoretical result.
  4. [§6.2, Table 3, Figure 9] The human evaluation is too thin to independently confirm the reported margin. Six annotators and 540 judgments are small, and no inter-annotator agreement (e.g., Fleiss' κ or ICC) is reported. The per-film correlations in Figure 9 use 90 points, but because the points come from only six systems, the effective sample for system-level inference is small; the reported p-values may be anti-conservative due to clustering. Please report the number of judgments per film per dimension, inter-annotator reliability, and cluster-aware or bootstrapped confidence intervals on the human ratings. This is important because the human evaluation is the only non-MLLM check on the central superiority claim.
minor comments (5)
  1. [Abstract, §5.3] The nine FilmEval metrics are called 'objective' in the abstract and Section 5.3, but they are scored by a prompted MLLM. 'Automatic' or 'LLM-judged' would be more accurate, and would avoid overclaiming objectivity relative to human perceptual judgment.
  2. [§6.2, Table 3] The text says six annotators rated 90 films on three dimensions, 'yielding 540 judgments.' The arithmetic is unclear: 90 × 3 × 6 = 1620, while 90 × 3 × 2 = 540. Please state explicitly how many judgments each annotator produced per film and per dimension.
  3. [References, §5.3] Reference [31] is cited for 'Gemini 3.1 Pro,' but the listed reference is the original Gemini 1.0/1.5 family paper (2023). Please cite the correct technical report or model card for the specific model version used, and verify all model-version citations in Section 6.1.
  4. [Table 2] The column header 'A VP' appears to be a typo for 'AVP' (Audio-Visual Performance Quality). Please correct for consistency with Table 1 and Section 5.3.
  5. [Appendix A.1, §E.2] Three benchmark entries (C4 from Conquer, C13 from Link Click, C15 from Scissor Seven) are 'screen adaptations' re-novelized from existing audiovisual works. The statement in Appendix A.1 that these were 'independently rewritten' and the claim in E.2 that FilmEval uses only 'original, public-domain, or substantially rewritten sources' may not be legally sufficient, since derivative works of copyrighted media can still infringe even after substantial rewriting. This is not a technical issue, but it should be clarified if the benchmark is released publicly.

Circularity Check

0 steps flagged

No significant circularity: the central derivation is self-contained, and the evaluation concerns are external-validity issues rather than constructional equivalence.

full rationale

The paper's load-bearing claims are (1) the analytical latency reduction in Section 4.4 and (2) the empirical superiority on FilmEval in Table 1. Neither reduces to its own inputs by construction. The O(N) to O(1) latency statement is an analytic consequence of the explicitly stated State Decoupling Condition (Eq. 10) — the paper itself notes the reduction holds only if the renderer's information set contains no rendered outputs from other shots, and FilmWorld's pre-materialized trajectory is designed to satisfy exactly that condition. This is a conditional mathematical claim, not a prediction that secretly re-imports its conclusion. The empirical comparison is measured by FilmEval, whose automatic metrics are all implemented by prompting Gemini 3.1 Pro (Section 5.3), the same model family used as FilmWorld's planner. This raises a legitimate evaluator-validity concern, but it is not circularity in the derivation sense: there is no equation or fitted parameter linking the generator's outputs to the judge's scores by construction. Moreover, the paper provides two forms of external check that break any same-family bias: Appendix B.4 re-runs the protocol with Qwen3.5-Omni and Seed2.0-Lite, with FilmWorld ranked first under all three judges, and Section 6.2 reports a human study of six annotators (540 judgments) whose system ordering matches the automatic ranking. Section E.1 candidly lists scope limitations (e.g., only 15 moderate-length narratives, no aesthetic-tone assessment), which are boundary statements rather than circular steps. No load-bearing self-citation was found: the references include prior agentic and story-generation systems, but the method's own contributions are argued from the formalization and ablations, not from a chain of same-author citations. Thus, under the rule that only constructional equivalence or fitted-input-renamed-as-prediction counts as circularity, the appropriate finding is no significant circularity, score 0.

Axiom & Free-Parameter Ledger

3 free parameters · 5 axioms · 1 invented entities

The formal model introduces no fitted constants, but the empirical pipeline relies on several hand-chosen hyperparameters (g=2, c=1, K=3) and on strong domain assumptions about LLM parsing, judge impartiality, and reference-following backbones. The invented entity is the world-state/identifier abstraction itself, which has no independent empirical handle.

free parameters (3)
  • sliding-window target shots g = 2
    Chosen by hand for the scene-local sliding-window planner; no sensitivity study is reported.
  • context shots c = 1
    Chosen by hand as padding around each sliding window; no sensitivity study is reported.
  • maximum correction rounds K = 3
    Chosen by hand; the diminishing-returns analysis in Section 6.3 justifies the choice but does not fit it.
axioms (5)
  • domain assumption LLM-based extraction from prose faithfully recovers the plot-event sequence and entity states.
    The entire pipeline treats the symbolic trajectory as ground truth; the paper acknowledges in Appendix E.1 that upstream entity-resolution and event-recovery errors can propagate.
  • domain assumption The MLLM judge (Gemini 3.1 Pro) provides an unbiased quantification of cinematic presentation, consistency, and novel fidelity.
    All nine FilmEval metrics are implemented as Gemini prompts (Section 5.3); human validation is limited to six annotators and does not eliminate model-bias risk.
  • domain assumption Rendering backbones honor reference assets and state identifiers.
    First-appearance visual anchoring assumes downstream generators condition on retrieved references; Figure 16 documents cases where they do not.
  • domain assumption The renderer's information set at shot t contains no rendered outputs of other shots (Eq. 10).
    This is the formal decoupling condition behind the O(1) latency claim; it holds by construction in FilmWorld but is an architectural assumption, not a theorem.
  • standard math Standard arithmetic and O-notation for latency summations.
    Equations 8-9 use conventional complexity arithmetic; no issue.
invented entities (1)
  • Dynamic Cinematic World state with state identifier Phi(W_t[e]) no independent evidence
    purpose: Central addressable representation mediating between abstract novel prose and visual rendering, claimed to prevent cross-shot drift.
    This is a design abstraction with no falsifiable prediction outside the paper; its utility is measured only through the paper's own benchmark and human study.

pith-pipeline@v1.3.0-alltime-deepseek · 3656 in / 5189 out tokens · 149115 ms · 2026-08-01T13:36:10.406008+00:00 · methodology

0 comments
read the original abstract

Translating novels into films poses a grand challenge for generative artificial intelligence, requiring conversion of abstract literary prose into long-form, multi-scene visual narratives. While current video generation models excel at short, single-scene clips within narrow temporal and spatial contexts, novel-to-film generation operates in a more complex regime, demanding long-duration content across diverse scenes with dynamically evolving entity states. To address this, we formalize novel-to-film generation as dynamic cinematic world modeling, decomposed into two phases: construction, which grounds abstract, underspecified literary narratives into concrete, stateful, and persistent world entities; and evolution, which governs how these entities dynamically update under plot progression to maintain causal consistency across scenes. We propose FilmWorld, an end-to-end agentic system where two groups of specialized agents collaborate to instantiate these phases. Construction-side agents perform narrative structured translation, world entity state modeling with visual anchoring, and state-driven shot planning, progressively projecting literary language into a cinematic blueprint. Evolution-side agents perform state-anchored visual generation, cross-shot dynamic state propagation, and closed-loop state verification to maintain causal consistency and visual coherence. To address the evaluation gap in long-form generation, we introduce FilmEval, a systematic evaluation framework that couples a difficulty-graded benchmark of 15 representative novels with an automated protocol of nine objective metrics spanning three dimensions: cinematic presentation, film consistency, and novel fidelity. Experiments demonstrate that FilmWorld consistently outperforms state-of-the-art video generation agent systems, with particularly pronounced improvements in narrative fidelity and cross-scene consistency.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

50 extracted references · 19 linked inside Pith

  1. [1]

    Onestory: Coherent multi-shot video generation with adaptive memory

    Zhaochong An et al. Onestory: Coherent multi-shot video generation with adaptive memory. InCVPR, 2026

    2026

  2. [2]

    Video generation models as world simulators

    Tim Brooks, Bill Peebles, Connor Holmes, Will DePue, Yufei Guo, Li Jing, David Schnurr, Joe Taylor, Troy Luhman, Eric Luhman, Clarence Ng, Ricky Wang, and Aditya Ramesh. Video generation models as world simulators. 2024. URL https://openai.com/index/video-generation-models-as-world-simulators/ . OpenAI Technical Report

    2024

  3. [3]

    Skyreels-v2: Infinite-length film generative model.arXiv preprint arXiv:2504.13074, 2025

    Guibin Chen, Dixuan Lin, Jiangping Yang, et al. Skyreels-v2: Infinite-length film generative model.arXiv preprint arXiv:2504.13074, 2025

    Pith/arXiv arXiv 2025

  4. [4]

    Self-forcing++: Towards minute-scale high-quality video generation

    Justin Cui, Jie Wu, Ming Li, Tao Yang, Xiaojie Li, Rui Wang, Andrew Bai, Yuanhao Ban, and Cho-Jui Hsieh. Self-forcing++: Towards minute-scale high-quality video generation. InICLR, 2026

    2026

  5. [5]

    A survey on long-video storytelling generation: Architectures, consistency, and cinematic quality.arXiv preprint arXiv:2507.07202, 2025

    Mohamed Elmoghany, Ryan Rossi, Seunghyun Yoon, et al. A survey on long-video storytelling generation: Architectures, consistency, and cinematic quality.arXiv preprint arXiv:2507.07202, 2025

    Pith/arXiv arXiv 2025

  6. [6]

    Infinitystory: Unlimited video generation with world consistency and character-aware shot transitions.arXiv preprint arXiv:2603.03646, 2026

    Mohamed Elmoghany, Liangbing Zhao, Xiaoqian Shen, Subhojyoti Mukherjee, Yang Zhou, Gang Wu, Viet Dac Lai, Seunghyun Yoon, Ryan Rossi, Abdullah Rashwan, et al. Infinitystory: Unlimited video generation with world consistency and character-aware shot transitions.arXiv preprint arXiv:2603.03646, 2026

    arXiv 2026

  7. [7]

    Talecrafter: Interactive story visualization with multiple characters

    Yuan Gong, Youxin Pang, Xiaodong Cun, Menghan Xia, Yingqing He, Haoxin Chen, Longyue Wang, Yong Zhang, Xintao Wang, Ying Shan, and Yujiu Yang. Talecrafter: Interactive story visualization with multiple characters. InSIGGRAPH Asia, 2023

    2023

  8. [8]

    Toonflow: Ai short drama factory, 2026

    HBAI Ltd. Toonflow: Ai short drama factory, 2026. URLhttps://github.com/HBAI-Ltd/Toonflow-app

    2026

  9. [9]

    Videoclaw: Ai creative video production agent, 2026

    HITsz-TMG. Videoclaw: Ai creative video production agent, 2026. URL https://github.com/HITsz-TMG/ VideoClaw

    2026

  10. [10]

    Storyagent: Customized storytelling video generation via multi-agent collaboration.arXiv preprint arXiv:2411.04925, 2024

    Panwen Hu, Jin Jiang, Jianqi Chen, Mingfei Han, Shengcai Liao, Xiaojun Chang, and Xiaodan Liang. Storyagent: Customized storytelling video generation via multi-agent collaboration.arXiv preprint arXiv:2411.04925, 2024

    Pith/arXiv arXiv 2024

  11. [11]

    Vabench: A comprehensive benchmark for audio-video generation.arXiv preprint arXiv:2512.09299, 2025

    Daili Hua, Xizhi Wang, Bohan Zeng, Xinyi Huang, Hao Liang, Junbo Niu, Xinlong Chen, Quanqing Xu, and Wentao Zhang. Vabench: A comprehensive benchmark for audio-video generation.arXiv preprint arXiv:2512.09299, 2025

    Pith/arXiv arXiv 2025

  12. [12]

    Filmaster: Bridging cinematic principles and generative ai for automated film generation.arXiv preprint arXiv:2506.18899, 2025

    Kaiyi Huang, Yukun Huang, Xintao Wang, Zinan Lin, Xuefei Ning, Pengfei Wan, Di Zhang, Yu Wang, and Xihui Liu. Filmaster: Bridging cinematic principles and generative ai for automated film generation.arXiv preprint arXiv:2506.18899, 2025

    Pith/arXiv arXiv 2025

  13. [13]

    Vimax: Agentic video generation.arXiv preprint arXiv:2606.07649, 2026

    Lingxuan Huang, Sizhe He, Hengji Zhou, Liqiang Nie, Lianghao Xia, and Chao Huang. Vimax: Agentic video generation.arXiv preprint arXiv:2606.07649, 2026

    Pith/arXiv arXiv 2026

  14. [14]

    Rehg, and Tobias Hinz

    Ozgur Kara, Krishna Kumar Singh, Feng Liu, Duygu Ceylan, James M. Rehg, and Tobias Hinz. Shotadapter: Text-to-multi-shot video generation with diffusion models. InCVPR, 2025

    2025

  15. [15]

    Hunyuanvideo: A systematic framework for large video generative models.arXiv preprint arXiv:2412.03603, 2024

    Weijie Kong, Qi Tian, Zijian Zhang, Rox Min, Zuozhuo Dai, Jin Zhou, Jiangfeng Xiong, Xin Li, Bo Wu, Jianwei Zhang, et al. Hunyuanvideo: A systematic framework for large video generative models.arXiv preprint arXiv:2412.03603, 2024

    Pith/arXiv arXiv 2024

  16. [16]

    Kling ai, 2024

    Kuaishou Technology. Kling ai, 2024. URLhttps://klingai.com/

    2024

  17. [17]

    Rolling sink: Bridging limited-horizon training and open-ended testing in autoregressive video diffusion.arXiv preprint arXiv:2602.07775, 2026

    Haodong Li, Shaoteng Liu, Zhe Lin, and Manmohan Chandraker. Rolling sink: Bridging limited-horizon training and open-ended testing in autoregressive video diffusion.arXiv preprint arXiv:2602.07775, 2026

    Pith/arXiv arXiv 2026

  18. [18]

    Storygan: A sequential conditional gan for story visualization

    Yitong Li, Zhe Gan, Yelong Shen, Jingjing Liu, Yu Cheng, Yuexin Wu, Lawrence Carin, David Carlson, and Jianfeng Gao. Storygan: A sequential conditional gan for story visualization. InCVPR, 2019

    2019

  19. [19]

    Videodirectorgpt: Consistent multi-scene video generation via llm-guided planning

    Han Lin, Abhay Zala, Jaemin Cho, and Mohit Bansal. Videodirectorgpt: Consistent multi-scene video generation via llm-guided planning. InCOLM, 2024

    2024

  20. [20]

    Reca: Multi-shot long video extrapolation via recursive context allocation.arXiv preprint arXiv:2605.26525, 2026

    Akide Liu, Jinbo Xing, Chaojie Mao, Ye Li, Zeyu Zhang, Yefei He, Weijie Wang, Zihan Wang, Yu Liu, Gholamreza Haffari, and Bohan Zhuang. Reca: Multi-shot long video extrapolation via recursive context allocation.arXiv preprint arXiv:2605.26525, 2026. 24

    Pith/arXiv arXiv 2026

  21. [21]

    Do Xuan Long, Xingchen Wan, Hootan Nakhost, Chen-Yu Lee, Tomas Pfister, and Sercan Ö. Arık. Vista: A test-time self-improving video generation agent.arXiv preprint arXiv:2510.15831, 2025

    arXiv 2025

  22. [22]

    Videostudio: Generating consistent-content and multi-scene videos

    Fuchen Long, Zhaofan Qiu, Ting Yao, and Tao Mei. Videostudio: Generating consistent-content and multi-scene videos. InECCV, 2024

    2024

  23. [23]

    Filmweaver: Weaving consistent multi-shot videos with cache-guided autoregressive diffusion

    Xiangyang Luo, Qingyu Li, Xiaokun Liu, Wenyu Qin, Miao Yang, Meng Wang, Pengfei Wan, Di Zhang, Kun Gai, and Shao-Lun Huang. Filmweaver: Weaving consistent multi-shot videos with cache-guided autoregressive diffusion. InAAAI, 2026

    2026

  24. [24]

    Storydall-e: Adapting pretrained text-to-image trans- formers for story continuation

    Adyasha Maharana, Darryl Hannan, and Mohit Bansal. Storydall-e: Adapting pretrained text-to-image trans- formers for story continuation. InECCV, 2022

    2022

  25. [25]

    Holocine: Holistic generation of cinematic multi-shot long video narratives

    Yihao Meng, Hao Ouyang, Yue Yu, Qiuyu Wang, Wen Wang, Ka Leong Cheng, Hanlin Wang, Shuailei Ma, Yixuan Li, Cheng Chen, Yanhong Zeng, Xing Zhu, Yujun Shen, and Huamin Qu. Holocine: Holistic generation of cinematic multi-shot long video narratives. InCVPR, 2026

    2026

  26. [26]

    The script is all you need: An agentic framework for long-horizon dialogue-to-cinematic video generation.arXiv preprint arXiv:2601.17737, 2026

    Chenyu Mu, Xin He, Qu Yang, Wanshun Chen, Jiadi Yao, Huang Liu, Zihao Yi, Bo Zhao, Xingyu Chen, Ruotian Ma, Fanghua Ye, Erkun Yang, Cheng Deng, Zhaopeng Tu, Xiaolong Li, and Linus. The script is all you need: An agentic framework for long-horizon dialogue-to-cinematic video generation.arXiv preprint arXiv:2601.17737, 2026

    Pith/arXiv arXiv 2026

  27. [27]

    Storygpt-v: Large language models as consistent story visualizers

    Xiaoqian Shen and Mohamed Elhoseiny. Storygpt-v: Large language models as consistent story visualizers. In CVPR, 2025

    2025

  28. [28]

    One sentence, one drama: Personalized short-form drama generation via multi-agent systems.arXiv preprint arXiv:2605.22144, 2026

    Yufei Shi, Weilong Yan, Naixuan Huang, Yucheng Chen, Chenyu Zhang, Tao He, Si Yong Yeo, and Ming Li. One sentence, one drama: Personalized short-form drama generation via multi-agent systems.arXiv preprint arXiv:2605.22144, 2026

    Pith/arXiv arXiv 2026

  29. [29]

    LeViet, Zack Chomyn, Joe Timmons, Brett Slatkin, Scott Penberthy, and Tomas Pfister

    Yale Song, Yiwen Song, Nick Losier, Nathan Hodson, Ye Jin, Rhyard Zhu, Yan Xu, Daniel Vlasic, Carina Claassen, Jasmine Leon, Khanh G. LeViet, Zack Chomyn, Joe Timmons, Brett Slatkin, Scott Penberthy, and Tomas Pfister. Co-director: Agentic generative video storytelling.arXiv preprint arXiv:2604.24842, 2026

    Pith/arXiv arXiv 2026

  30. [30]

    Muse: A multi-agent frame- work for unconstrained story envisioning via closed-loop cognitive orchestration.arXiv preprint arXiv:2602.03028, 2026

    Wenzhang Sun, Zhenyu Wang, Zhangchi Hu, Chunfeng Wang, Hao Li, and Wei Chen. Muse: A multi-agent frame- work for unconstrained story envisioning via closed-loop cognitive orchestration.arXiv preprint arXiv:2602.03028, 2026

    arXiv 2026

  31. [31]

    Gemini: a family of highly capable multimodal models.arXiv preprint arXiv:2312.11805, 2023

    Gemini Team, Rohan Anil, Sebastian Borgeaud, Jean-Baptiste Alayrac, Jiahui Yu, Radu Soricut, Johan Schalkwyk, Andrew M Dai, Anja Hauth, Katie Millican, et al. Gemini: a family of highly capable multimodal models.arXiv preprint arXiv:2312.11805, 2023

    Pith/arXiv arXiv 2023

  32. [32]

    Wan: Open and advanced large-scale video generative models.arXiv preprint arXiv:2503.20314, 2025

    Team Wan, Ang Wang, Baole Ai, Bin Wen, Chaojie Mao, Chen-Wei Xie, Di Chen, Feiwu Yu, Haiming Zhao, Jianxiao Yang, Jianyuan Zeng, et al. Wan: Open and advanced large-scale video generative models.arXiv preprint arXiv:2503.20314, 2025

    Pith/arXiv arXiv 2025

  33. [33]

    Huang, Hyeonho Jeong, and Yulia Gryaditskaya

    Jente Vandersanden, Matheus Gadelha, Chun-Hao P. Huang, Hyeonho Jeong, and Yulia Gryaditskaya. Em-vid: Training-free entity-centric memory for efficient and consistent multi-shot video generation.arXiv preprint arXiv:2605.23610, 2026

    Pith/arXiv arXiv 2026

  34. [34]

    Phenaki: Variable length video generation from open domain textual descriptions

    Ruben Villegas, Mohammad Babaeizadeh, Pieter-Jan Kindermans, Hernan Moraldo, Han Zhang, Mohammad Taghi Saffar, Santiago Castro, Julius Kunze, and Dumitru Erhan. Phenaki: Variable length video generation from open domain textual descriptions. InICLR, 2023

    2023

  35. [35]

    Mavis: A multi-agent framework for long-sequence video storytelling

    Qian Wang, Ziqi Huang, Ruoxi Jia, Paul Debevec, and Ning Yu. Mavis: A multi-agent framework for long-sequence video storytelling. InEACL, 2026

    2026

  36. [36]

    Msavbench: Towards comprehensive and reliable evaluation of multi-shot audio-video generation.arXiv preprint arXiv:2605.20183, 2026

    Yujie Wei, Yujin Han, Zhekai Chen, Yongming Li, Kaixun Jiang, Zhihang Liu, Quanhao Li, Zhiwu Qing, Xiang Wang, Zhen Xing, et al. Msavbench: Towards comprehensive and reliable evaluation of multi-shot audio-video generation.arXiv preprint arXiv:2605.20183, 2026

    Pith/arXiv arXiv 2026

  37. [37]

    Moviebench: A hierarchical movie level dataset for long video generation

    Weijia Wu, Mingyu Liu, Zeyu Zhu, Xi Xia, Haoen Feng, Wen Wang, Kevin Qinghong Lin, Chunhua Shen, and Mike Zheng Shou. Moviebench: A hierarchical movie level dataset for long video generation. InCVPR, 2025

    2025

  38. [38]

    Automated movie generation via multi-agent cot planning.arXiv preprint arXiv:2503.07314, 2025

    Weijia Wu, Zeyu Zhu, and Mike Zheng Shou. Automated movie generation via multi-agent cot planning.arXiv preprint arXiv:2503.07314, 2025

    Pith/arXiv arXiv 2025

  39. [39]

    Large multimodal agents: A survey.Visual Intelligence, 3(24), 2025

    Junlin Xie, Zhihong Chen, Ruifei Zhang, Xiang Wan, and Guanbin Li. Large multimodal agents: A survey.Visual Intelligence, 3(24), 2025. 25

    2025

  40. [40]

    Mm-storyagent: Immersive narrated storybook video generation with a multi-agent paradigm across text, image and audio.arXiv preprint arXiv:2503.05242, 2025

    Xuenan Xu, Jiahao Mei, Chenliang Li, Yuning Wu, Ming Yan, Shaopeng Lai, Ji Zhang, and Mengyue Wu. Mm-storyagent: Immersive narrated storybook video generation with a multi-agent paradigm across text, image and audio.arXiv preprint arXiv:2503.05242, 2025

    Pith/arXiv arXiv 2025

  41. [41]

    Evalverse: Pipeline-aware and expert-calibrated benchmarking for professional cinematic video generation.arXiv preprint arXiv:2605.23271, 2026

    Songlin Yang, Haobin Zhong, Ruilin Zhang, et al. Evalverse: Pipeline-aware and expert-calibrated benchmarking for professional cinematic video generation.arXiv preprint arXiv:2605.23271, 2026

    Pith/arXiv arXiv 2026

  42. [42]

    Cogvideox: Text-to-video diffusion models with an expert transformer

    Zhuoyi Yang, Jiayan Teng, Wendi Zheng, Ming Ding, Shiyu Huang, Jiazheng Xu, Yuanming Yang, Wenyi Hong, Xiaohan Zhang, Guanyu Feng, et al. Cogvideox: Text-to-video diffusion models with an expert transformer. In ICLR, 2025

    2025

  43. [43]

    Narrative weaver: Towards controllable long-range visual consistency with multi-modal conditioning

    Zhengjian Yao, Yongzhi Li, Xinyuan Gao, Quan Chen, Peng Jiang, and Yanye Lu. Narrative weaver: Towards controllable long-range visual consistency with multi-modal conditioning. InCVPR, 2026

    2026

  44. [44]

    Storymem: Multi-shot long video storytelling with memory.arXiv preprint arXiv:2512.19539, 2025

    Kaiwen Zhang et al. Storymem: Multi-shot long video storytelling with memory.arXiv preprint arXiv:2512.19539, 2025

    arXiv 2025

  45. [45]

    Anime: Adaptive multi-agent planning for long animation generation

    Lisai Zhang, Baohan Xu, Siqian Yang, Mingyu Yin, Jing Liu, Chao Xu, Siqi Wang, Yidi Wu, Yuxin Hong, Zihao Zhang, Yanzhang Liang, and Yudong Jiang. Anime: Adaptive multi-agent planning for long animation generation. InSIGGRAPH Asia, 2025

    2025

  46. [46]

    Stage: Storyboard-anchored generation for cinematic multi-shot narrative.arXiv preprint arXiv:2512.12372, 2025

    Peixuan Zhang, Zijian Jia, Kaiqi Liu, Shuchen Weng, Si Li, and Boxin Shi. Stage: Storyboard-anchored generation for cinematic multi-shot narrative.arXiv preprint arXiv:2512.12372, 2025

    arXiv 2025

  47. [47]

    Videogen-of-thought: Step-by-step generating multi-shot video with minimal manual intervention

    Mingzhe Zheng, Yongqi Xu, Haojian Huang, Xuran Ma, Yexin Liu, Wenjie Shu, Yatian Pang, Feilong Tang, Qifeng Chen, Harry Yang, and Ser-Nam Lim. Videogen-of-thought: Step-by-step generating multi-shot video with minimal manual intervention. InNeurIPS NextVid Workshop, 2025

    2025

  48. [48]

    Videomemory: Toward consistent video generation via memory integration.arXiv preprint arXiv:2601.03655, 2026

    Jinsong Zhou, Yihua Du, Xinli Xu, Luozhou Wang, Zijie Zhuang, Yehang Zhang, Shuaibo Li, Xiaojun Hu, Bolan Su, and Ying-cong Chen. Videomemory: Toward consistent video generation via memory integration.arXiv preprint arXiv:2601.03655, 2026

    arXiv 2026

  49. [49]

    Storydiffusion: Consistent self-attention for long-range image and video generation

    Yupeng Zhou, Daquan Zhou, Ming-Ming Cheng, Jiashi Feng, and Qibin Hou. Storydiffusion: Consistent self-attention for long-range image and video generation. InNeurIPS, 2024

    2024

  50. [50]

    I failed my exam today and don’t want to go back to my dorm

    Cailin Zhuang, Ailin Huang, Yaoqi Hu, Jingwei Wu, Wei Cheng, Jiaqi Liao, Hongyuan Wang, Xinyao Liao, Weiwei Cai, Hengyuan Xu, Xuanyang Zhang, Xianfang Zeng, Zhewei Huang, Gang Yu, and Chi Zhang. Vistorybench: Comprehensive benchmark suite for story visualization. InCVPR, 2026. 26 Appendix A FilmEval Benchmark Details This appendix expands the one-paragrap...

    2026