Pith. sign in

REVIEW 4 major objections 6 minor 98 references

VerIPO: Cultivating Long Reasoning in Video-LLMs via Verifier-Gudied Iterative Policy Optimization

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Verifier-curated preference pairs can replace long-CoT cold starts for video reasoning, and one iteration of VerIPO already beats much larger models.

desk verdict A useful training recipe with credible benchmark gains, but the consistency metric is self-referential and the 'reasoning logic' claim is not backed by the measurements. read the letter →

arxiv 2505.19000 v1 pith:GUNOICWV submitted 2025-05-25 cs.CL cs.CV

Yunxin Li , Xinyu Chen , Zitao Li , Zhenyu Liu , Longyue Wang , Wenhan Luo , Baotian Hu , Min Zhang This is my paper · ORCID
classification cs.CLcs.CV
keywords videoreasoninglongchain-of-thoughtreinforcementlearningGRPODPOpreferenceoptimizationverifier-guidedtrainingmultimodalLLM
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

VerIPO claims that a Video-LLM's long reasoning ability can be cultivated without the expensive, noisy cold start of supervised fine-tuning on long chain-of-thought (CoT) datasets. The method alternates GRPO's broad search over answer rollouts with a verifier stage that curates preference pairs, and a DPO stage that aligns the policy to those pairs. Across video reasoning and long-video benchmarks, the paper reports consistent gains in accuracy, reasoning length, and consistency between the thinking and final answer, with one iteration already outperforming much larger instruction-tuned models and existing RL-trained long-reasoning models. If correct, the contribution is a cheaper and more stable route to activating long reasoning in video models.

What carries the argument

The load-bearing mechanism is the Rollout-Aware Verifier, a small language-model judge paired with rule-based checks, placed between GRPO and DPO. It extracts the chain-of-thought and final answer from each rollout, checks whether the extracted reasoning answer matches the predicted answer, flags repetitions, and prefers longer verified-correct chains; these labels are then turned into four categories of preference pairs that drive DPO. The verifier also prunes already-mastered simple samples to speed up later iterations, and a preceding reasoning-activation stage using text and image math data seeds the model's ability to explore long logical paths before video GRPO begins.

What would settle it

Take a held-out set of video questions with human labels for whether the chain of thought truly supports the final answer, run the verifier's consistency check on fresh rollouts, and measure agreement with the human labels. If agreement is low on hard questions while VerIPO's reported inconsistency rate drops, the gains are partly fitting the verifier rather than improving reasoning.

Watch

Extended reading notes

Core claim

The central claim is that placing a rollout-aware verifier between GRPO and DPO turns unstable outcome-based RL into a stable iterative training loop for long video reasoning. The verifier scores each rollout on accuracy, consistency between the extracted reasoning and the final answer, repetition, and length, and from those scores builds four kinds of preference pairs: single-turn pairs favouring verified correct long rollouts, repetition-penalty pairs, reflective pairs that concatenate wrong tries followed by a correct try, and inference-consistency pairs that reject rollouts whose thinking contradicts the answer. Training with DPO on these pairs is about seven times faster per sample than GRPO and, iterated, it increases true accuracy, response length, and reasoning-answer consistency while reducing repetition. The paper reports that one VerIPO iteration on a 7B video model surpasses larger direct-answer models and long-reasoning RL baselines on spatial, temporal, and multi-discipline video benchmarks, with further iterations giving stable additional gains.

Load-bearing premise

The method assumes the verifier's consistency check—a small model extracting a reasoning answer and a rule-based matcher comparing it with the predicted answer—is accurate enough that the preference data it generates is genuinely better, not just better at pleasing the verifier.

Editorial extensions

If this is right

  • Verifier-guided DPO is about 7x faster per training sample than GRPO, so the loop cuts the compute cost of RL fine-tuning for video reasoning.
  • Iterating the loop raises true accuracy (correct answer with consistent reasoning) and lowers reasoning-answer inconsistency rates, from double digits to the low single digits on multiple benchmarks.
  • The method removes the need for a large long-CoT SFT cold start; in the paper's comparison, the SFT cold start degraded general video benchmarks and was not recovered by later VerIPO iterations.
  • Later iterations trained only on video data continue to improve accuracy and chain length, so the approach can scale by adding harder video questions rather than more long-CoT annotations.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Inference: the verifier is the real bottleneck, because DPO amplifies whatever signal the preference pairs carry; a systematically biased consistency judge would bake its bias into the policy, so the method's ceiling is set by the judge's reliability.
  • Inference: the same GRPO-verifier-DPO loop should transfer to other domains where answers are verifiable but reasoning quality matters, such as math, code, or agentic planning, provided a suitable consistency judge exists.
  • Inference: a direct test of the mechanism is to train with human-annotated reasoning-consistency labels instead of the verifier's labels and compare final inconsistency rates; high agreement would confirm the verifier is learning real reasoning rather than mimicking the judge.
  • Inference: the loop could be extended with adaptive-reasoning rewards that penalize overthinking on simple questions, since the paper's own discussion identifies overthinking as a weakness of long-reasoning models.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper proposes VerIPO, an iterative training procedure for Video-LLMs that interleaves GRPO with a verifier-guided DPO stage. A rollout-aware verifier scores rollouts on answer accuracy, reasoning-answer consistency, repetition, and length, then constructs preference pairs (single-turn, repetition-penalty, reflective, and inference-consistency pairs) for a DPO update. Experiments on VSI-Bench, Video-MMMU, MMVU, TOMATO, LVBench, and Video-MME report consistent gains in accuracy and response length, with one iteration of VerIPO outperforming larger instruction-tuned models and long-reasoning models such as Kimi-VL-Thinking and Video-R1.

Significance. The central idea---avoiding large long-CoT SFT cold starts by using a verifier to curate preference pairs from GRPO rollouts and then applying fast DPO---is timely and practically motivated. The external benchmark accuracies in Table 2 are grounded in ground truth and do not appear circular, and the reported 7x training-speed advantage of DPO over GRPO (Section A.1) is a useful practical result. The paper also ships code, which supports reproducibility. However, the paper's strongest qualitative claim, that the verifier 'assesses reasoning logic' and yields contextually consistent chains, is supported only by metrics derived from the same Qwen3-8B extraction and rule-based matching used to build the training signal. The causal story therefore needs additional, independent validation before the method's contribution can be fully accepted.

major comments (4)
  1. [Section 4.3, Consistency Check] The Consistency Check only compares a Qwen3-8B-extracted reasoning answer against the final answer using a rule-based matcher; it does not assess whether the reasoning itself is logically correct. A wrong derivation that ends with the correct answer passes the check. Consequently, the 'contextually consistent CoTs' used as DPO positives are not actually verified for reasoning quality, and the reported improvements in consistency (Table 6, Figure 3B, Acc-True in Table 3) could reflect the policy learning to echo the final answer inside <think> rather than genuine reasoning gains. A human- or stronger-judge agreement study on a sample of rollouts, reported separately for chosen versus unchosen reasoning paths, would settle whether the verifier measures logical consistency.
  2. [Section 4.3 and Table 3 / Figure 3B] The evaluation metrics Acc-True and inconsistency rate use the same Qwen3-8B extraction and rule-based matching that the Verifier uses to construct the training signal. Thus the improvement in these metrics is partly circular: the DPO stage directly rewards rollouts that pass this exact check. The external accuracies in Table 2 are not affected by this circularity, but the specific claim that VerIPO improves 'contextual consistency' in Section 5.2 rests on a metric that is not independent of the training objective. I recommend reporting at least one external consistency metric (e.g., a stronger LLM judge or human annotation) to break the circularity.
  3. [Table 2 and Section A.5] Table 2 reports many baseline numbers without stating which were re-evaluated under the authors' evaluation protocol (2.0 fps, up to 128 frames, 256x28x28 resolution). Numbers copied from prior papers may not be directly comparable, especially for long-video benchmarks where sampling density matters: Table 8 shows that changes in frames and resolution alter MMVU (mc) by up to 5 points. Please mark which baselines were re-run and which are cited from original papers, and justify comparability given the evaluation settings.
  4. [Table 3 and Figure 3] All training curves and ablation results are single-run with no error bars or multiple seeds. Given the known instability of GRPO-based training acknowledged in Section 1, and the small differences between iterations (e.g., VSI-Bench 41.8, 41.0, 41.3 in Table 2), the reported trends are not statistically assessable. Reporting at least two or three seeds for the key comparison (VerIPO vs. GRPO-only) would substantially strengthen the claim of stable improvement.
minor comments (6)
  1. [Title] The title contains a typo: 'Verifier-Gudied' should be 'Verifier-Guided'.
  2. [Table 2] Several cells contain dashes with no explanation; this makes row-to-row comparison difficult. Please add a footnote clarifying whether '-' means 'not reported' or 'not applicable'.
  3. [Figure 3(C)] The caption says 'number of repeated responses' but the text says the values are computed as a sum across four datasets. Please clarify the exact metric and the unit (responses, ratio, or count).
  4. [Section A.2] Table 6 reports inconsistency rates but the formal definition of the inconsistency rate is not given. Please provide an equation or explicit algorithmic description so readers can replicate the metric.
  5. [Section 4.3] The MRA correctness threshold of 0.6 for distance estimation tasks is introduced without justification or sensitivity analysis. Please discuss the choice and its effect on the constructed preference data.
  6. [Section 5.1] The statement that video training data was 'carefully checked for leakage with evaluation datasets' is not accompanied by any detail. Please specify the filtering procedure and report overlap statistics with the evaluation benchmarks.

Circularity Check

1 steps flagged · score 4.0 of 10

Consistency metric is self-referential; external benchmark gains remain independent.

  1. fitted input called prediction [Section 4.3 (Consistency Check; Inference Consistency Pairs), Section 5.2 / Table 3 / Table 6 / Figure 3B, A.2]
    "Consistency Check evaluates the coherence between the reasoning process and the final answer. It uses the Qwen3-8B to extract the reasoning answer a_i^r from the response r_i, given the original question. A rule-based method then verifies if a_i^r matches the predicted answer a_i to assess reasoning consistency. ... Rollouts with correct reasoning but incorrect answers are rejected against a chosen sample using the correct reasoning path with the extracted reasoning answer (a_i^r). ... “Acc-True” indicates the answer is correct and also consistent with the right reasoning process."

    DPO training uses exactly the Consistency Check as its preference label: chosen rollouts are the longest ones that passed this check, and Inference Consistency Pairs reject rollouts whose <think> and <answer> disagree, even rewriting the final answer to the extracted thought answer for the chosen sample. The reported evidence for improved consistency (Table 6, Figure 3B, Acc-True in Table 3) is computed with the same Qwen3-8B extraction + rule-match protocol. A model can reduce this metric by restating its final answer inside <think>, which is precisely what the DPO positives reward. The MMVU inconsistency drop from 13.3 to 2.1 is therefore partly fitting the verifier's operational rule, not an independent measure of reasoning logic. Table 2 external-GT accuracy gains remain non-circular.

full rationale

Most headline accuracy comparisons (Table 2: VSI-Bench, Video-MMMU, MMVU, TOMATO, LVBench, Video-MME) are scored against external benchmark ground truths, so the main empirical gains are not circular. No load-bearing self-citations are used; references to DeepSeek-R1, DAPO, and standard RL methods are contextual. The circularity is confined to the internal consistency/Acc-True metrics: the Consistency Check operationally defines consistency as the Qwen3-8B-extracted thought answer matching the final answer; the DPO positives and inference-consistency pairs are built from that exact function; and the evaluation tables report that same function. The “contextually consistent CoTs” and “true accuracy” improvements are therefore partly self-referential, though the external accuracy improvements and length statistics are independently meaningful. A human-agreement or logical-validity evaluation of the CoTs would be needed to support the claim that the verifier “assesses reasoning logic.” Score 4 reflects partial circularity in one internal metric, with the external benchmark claims remaining independent.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claims rest on the reliability of the verifier's quality checks, the absence of data leakage, and several hand-chosen thresholds. Accuracy gains are anchored to external benchmark ground truths, but consistency and length claims are partly defined by the same verifier signals used to build training data.

free parameters (4)
  • MRA correctness threshold = 0.6
    In Section 4.3, a distance-estimation rollout is treated as correct only if MRA exceeds 0.6; no sensitivity analysis is provided.
  • DPO beta = 0.1
    DPO temperature hyperparameter chosen in Table 1 with no reported sweep.
  • Positive-sample length preference = longest verified correct rollout
    Section 4.3 selects the longest verified correct rollout as the chosen example, so response-length gains are built into the training signal.
  • Simple-data discard rate = 80%
    Section 4.4 discards 80% of simple examples with perfect rollout accuracy each iteration, changing the training distribution.
assumptions (5)
  • domain assumption Qwen3-8B can reliably extract the reasoning answer from CoT text for consistency checking
    Section 4.3 Consistency Check makes this assumption without human agreement or error analysis.
  • domain assumption Training data is free of leakage with the evaluation benchmarks
    Section 5.1 says the filtered video subset was carefully checked for leakage but gives no procedure or evidence.
  • domain assumption Removing the KL penalty in GRPO keeps training stable
    Section 4.2 removes KL divergence and cites DAPO; no ablation for this choice is reported.
  • domain assumption Benchmark ground truths and accuracy reward matching are reliable
    Reward design in Section 4.2 assumes exact match for MC and math questions and MRA for distance estimation, and these labels feed the verifier.
  • ad hoc to paper Longer verified reasoning is better for downstream tasks
    Positive samples are selected by length among correct rollouts, treating length as a quality signal.

how reviews work

0 comments
Cite this review

Pith. "Pith review of VerIPO: Cultivating Long Reasoning in Video-LLMs via Verifier-Gudied Iterative Policy Optimization." pith.science (2026). https://pith.science/paper/GUNOICWV

@misc{pith2026250519000,
  author       = {Pith},
  title        = {Pith review of: VerIPO: Cultivating Long Reasoning in Video-LLMs via Verifier-Gudied Iterative Policy Optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GUNOICWV}},
  note         = {Machine review of arXiv:2505.19000}
}
read the original abstract

Applying Reinforcement Learning (RL) to Video Large Language Models (Video-LLMs) shows significant promise for complex video reasoning. However, popular Reinforcement Fine-Tuning (RFT) methods, such as outcome-based Group Relative Policy Optimization (GRPO), are limited by data preparation bottlenecks (e.g., noise or high cost) and exhibit unstable improvements in the quality of long chain-of-thoughts (CoTs) and downstream performance.To address these limitations, we propose VerIPO, a Verifier-guided Iterative Policy Optimization method designed to gradually improve video LLMs' capacity for generating deep, long-term reasoning chains. The core component is Rollout-Aware Verifier, positioned between the GRPO and Direct Preference Optimization (DPO) training phases to form the GRPO-Verifier-DPO training loop. This verifier leverages small LLMs as a judge to assess the reasoning logic of rollouts, enabling the construction of high-quality contrastive data, including reflective and contextually consistent CoTs. These curated preference samples drive the efficient DPO stage (7x faster than GRPO), leading to marked improvements in reasoning chain quality, especially in terms of length and contextual consistency. This training loop benefits from GRPO's expansive search and DPO's targeted optimization. Experimental results demonstrate: 1) Significantly faster and more effective optimization compared to standard GRPO variants, yielding superior performance; 2) Our trained models exceed the direct inference of large-scale instruction-tuned Video-LLMs, producing long and contextually consistent CoTs on diverse video reasoning tasks; and 3) Our model with one iteration outperforms powerful LMMs (e.g., Kimi-VL) and long reasoning models (e.g., Video-R1), highlighting its effectiveness and stability.

Figures

Figures reproduced from arXiv: 2505.19000 by the authors.

Figure 1
Figure 1. Figures (A, D): Initial GRPO training with different data types shows only utilizing [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Overview of VerIPO workflow. This training loop is guided by the Verifier’s continuous [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure (A): Performance comparison after removing [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: A case from Video-MMMU shows the comparative performance of GRPO and VerIPO. [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: A case from VSI-Bench shows the comparative performance of GRPO and VerIPO. Our [PITH_FULL_IMAGE:figures/full_fig_p018_5.png]
Figure 6
Figure 6. Figure 6: Another case from VSI-Bench shows the comparative performance of GRPO and VerIPO. [PITH_FULL_IMAGE:figures/full_fig_p018_6.png]
Figure 7
Figure 7. Figure 7: A case from Video-MMMU shows the comparative performance of GRPO and VerIPO. [PITH_FULL_IMAGE:figures/full_fig_p018_7.png]
Figure 8
Figure 8. Figure 8: A case from Video-MME shows the comparative performance of GRPO and VerIPO. [PITH_FULL_IMAGE:figures/full_fig_p019_8.png]
Figure 9
Figure 9. Figure 9: A case from TOMATO shows the comparative performance of GRPO and VerIPO. Our [PITH_FULL_IMAGE:figures/full_fig_p019_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

98 extracted references · 38 canonical work pages

  1. [1]

    Vivit: A video vision transformer, 2021

    Anurag Arnab, Mostafa Dehghani, Georg Heigold, Chen Sun, Mario Lu ˇci´c, and Cordelia Schmid. Vivit: A video vision transformer, 2021

    2021

  2. [2]

    Qwen2.5-vl technical report, 2025

    Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, Humen Zhong, Yuanzhi Zhu, Mingkun Yang, Zhaohai Li, Jianqiang Wan, Pengfei Wang, Wei Ding, Zheren Fu, Yiheng Xu, Jiabo Ye, Xi Zhang, Tianbao Xie, Zesen Cheng, Hang Zhang, Zhibo Yang, Haiyang Xu, and Junyang Lin. Qwen2.5-vl technical report, 2025

    2025

  3. [3]

    Hadzic, Taran Kota, Jimming He, Cristobal Eyzaguirre, Zane Durante, Manling Li, Jiajun Wu, and Fei-Fei Li

    Keshigeyan Chandrasegaran, Agrim Gupta, Lea M. Hadzic, Taran Kota, Jimming He, Cristobal Eyzaguirre, Zane Durante, Manling Li, Jiajun Wu, and Fei-Fei Li. Hourvideo: 1-hour video- language understanding. In Advances in Neural Information Processing Systems, volume 37, 2024

    2024

  4. [4]

    Mecd: Unlocking multi-event causal discovery in video reasoning, 2024

    Tieyuan Chen, Huabin Liu, Tianyao He, Yihang Chen, Chaofan Gan, Xiao Ma, Cheng Zhong, Yang Zhang, Yingxue Wang, Hui Lin, and Weiyao Lin. Mecd: Unlocking multi-event causal discovery in video reasoning, 2024

    2024

  5. [5]

    On the suitability of reinforcement fine-tuning to visual tasks, 2025

    Xiaxu Chen, Wei Li, Chunxu Liu, Chi Xie, Xiaoyan Hu, Chengqian Ma, Feng Zhu, and Rui Zhao. On the suitability of reinforcement fine-tuning to visual tasks, 2025

    2025

  6. [6]

    Videovista-culturallingo: 360◦ horizons-bridging cultures, languages, and domains in video comprehension, 2025

    Xinyu Chen, Yunxin Li, Haoyuan Shi, Baotian Hu, Wenhan Luo, Yaowei Wang, and Min Zhang. Videovista-culturallingo: 360◦ horizons-bridging cultures, languages, and domains in video comprehension, 2025

    2025

  7. [7]

    Visrl: Intention-driven visual perception via reinforced reasoning, 2025

    Zhangquan Chen, Xufang Luo, and Dongsheng Li. Visrl: Intention-driven visual perception via reinforced reasoning, 2025

    2025

  8. [8]

    Expanding performance boundaries of open-source multimodal models with model, data, and test-time scaling, 2025

    Zhe Chen, Weiyun Wang, Yue Cao, Yangzhou Liu, Zhangwei Gao, Erfei Cui, Jinguo Zhu, Shenglong Ye, Hao Tian, Zhaoyang Liu, Lixin Gu, Xuehui Wang, Qingyun Li, Yimin Ren, Zixuan Chen, Jiapeng Luo, Jiahao Wang, Tan Jiang, Bo Wang, Conghui He, Botian Shi, Xingcheng Zhang, Han Lv, Yi Wang, Wenqi Shao, Pei Chu, Zhongying Tu, Tong He, Zhiyong Wu, Huipeng Deng, Jia...

    2025

Show all 98 references
  1. [9]

    Videollama 2: Advancing spatial- temporal modeling and audio understanding in video-llms, 2024

    Zesen Cheng, Sicong Leng, Hang Zhang, Yifei Xin, Xin Li, Guanzheng Chen, Yongxin Zhu, Wenqi Zhang, Ziyang Luo, Deli Zhao, and Lidong Bing. Videollama 2: Advancing spatial- temporal modeling and audio understanding in video-llms, 2024

    2024

  2. [10]

    Skywork r1v2: Multimodal hybrid reinforcement learning for reasoning, 2025

    Chris, Yichen Wei, Yi Peng, Xiaokun Wang, Weijie Qiu, Wei Shen, Tianyidan Xie, Jiangbo Pei, Jianhao Zhang, Yunzhuo Hao, Xuchen Song, Yang Liu, and Yahui Zhou. Skywork r1v2: Multimodal hybrid reinforcement learning for reasoning, 2025

    2025

  3. [11]

    Instructblip: Towards general-purpose vision-language models with instruction tuning, 2023

    Wenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong, Junqi Zhao, Weisheng Wang, Boyang Li, Pascale Fung, and Steven Hoi. Instructblip: Towards general-purpose vision-language models with instruction tuning, 2023

    2023

  4. [12]

    Mm-spatial: Exploring 3d spatial understanding in multimodal llms, 2025

    Erik Daxberger, Nina Wenzel, David Griffiths, Haiming Gang, Justin Lazarow, Gefen Kohavi, Kai Kang, Marcin Eichner, Yinfei Yang, Afshin Dehghan, and Peter Grasch. Mm-spatial: Exploring 3d spatial understanding in multimodal llms, 2025

    2025

  5. [13]

    Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning, 2025

    DeepSeek-AI, Daya Guo, Dejian Yang, Haowei Zhang, et al. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning, 2025. 11

    2025

  6. [14]

    Boosting the generalization and reasoning of vision language models with curriculum reinforcement learning, 2025

    Huilin Deng, Ding Zou, Rui Ma, Hongchen Luo, Yang Cao, and Yu Kang. Boosting the generalization and reasoning of vision language models with curriculum reinforcement learning, 2025

    2025

  7. [15]

    Insight-v: Exploring long-chain visual reasoning with multimodal large language models, 2025

    Yuhao Dong, Zuyan Liu, Hai-Long Sun, Jingkang Yang, Winston Hu, Yongming Rao, and Ziwei Liu. Insight-v: Exploring long-chain visual reasoning with multimodal large language models, 2025

    2025

  8. [16]

    An image is worth 16x16 words: Transformers for image recognition at scale, 2021

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition at...

    2021

  9. [17]

    Video-of-thought: Step-by-step video reasoning from perception to cognition, 2024

    Hao Fei, Shengqiong Wu, Wei Ji, Hanwang Zhang, Meishan Zhang, Mong-Li Lee, and Wynne Hsu. Video-of-thought: Step-by-step video reasoning from perception to cognition, 2024

    2024

  10. [18]

    Video-r1: Reinforcing video reasoning in mllms, 2025

    Kaituo Feng, Kaixiong Gong, Bohao Li, Zonghao Guo, Yibing Wang, Tianshuo Peng, Benyou Wang, and Xiangyu Yue. Video-r1: Reinforcing video reasoning in mllms, 2025

    2025

  11. [19]

    Video-mme: The first-ever comprehensive evaluation benchmark of multi-modal llms in video analysis

    Chaoyou Fu, Yuhan Dai, Yondong Luo, Lei Li, Shuhuai Ren, Renrui Zhang, Zihan Wang, Chenyu Zhou, Yunhang Shen, Mengdan Zhang, et al. Video-mme: The first-ever comprehensive evaluation benchmark of multi-modal llms in video analysis. arXiv preprint arXiv:2405.21075, 2024

    2024 arXiv

  12. [20]

    Ampo: Active multi-preference optimization, 2025

    Taneesh Gupta, Rahul Madhavan, Xuchao Zhang, Chetan Bansal, and Saravan Rajmohan. Ampo: Active multi-preference optimization, 2025

    2025

  13. [21]

    Video-mmmu: Evaluating knowledge acquisition from multi-discipline professional videos, 2025

    Kairui Hu, Penghao Wu, Fanyi Pu, Wang Xiao, Yuanhan Zhang, Xiang Yue, Bo Li, and Ziwei Liu. Video-mmmu: Evaluating knowledge acquisition from multi-discipline professional videos, 2025

    2025

  14. [22]

    Vision-r1: Incentivizing reasoning capability in multimodal large language models, 2025

    Wenxuan Huang, Bohan Jia, Zijie Zhai, Shaosheng Cao, Zheyu Ye, Fei Zhao, Zhe Xu, Yao Hu, and Shaohui Lin. Vision-r1: Incentivizing reasoning capability in multimodal large language models, 2025

    2025

  15. [23]

    Gpt-4o system card

    Aaron Hurst, Adam Lerer, Adam P Goucher, Adam Perelman, Aditya Ramesh, Aidan Clark, AJ Ostrow, Akila Welihinda, Alan Hayes, Alec Radford, et al. Gpt-4o system card. arXiv preprint arXiv:2410.21276, 2024

    2024 arXiv

  16. [24]

    Llava-onevision: Easy visual task transfer, 2024

    Bo Li, Yuanhan Zhang, Dong Guo, Renrui Zhang, Feng Li, Hao Zhang, Kaichen Zhang, Peiyuan Zhang, Yanwei Li, Ziwei Liu, and Chunyuan Li. Llava-onevision: Easy visual task transfer, 2024

    2024

  17. [25]

    mplug: Effective and efficient vision-language learning by cross-modal skip-connections, 2022

    Chenliang Li, Haiyang Xu, Junfeng Tian, Wei Wang, Ming Yan, Bin Bi, Jiabo Ye, Hehong Chen, Guohai Xu, Zheng Cao, Ji Zhang, Songfang Huang, Fei Huang, Jingren Zhou, and Luo Si. mplug: Effective and efficient vision-language learning by cross-modal skip-connections, 2022

    2022

  18. [26]

    Videochat: Chat-centric video understanding, 2024

    KunChang Li, Yinan He, Yi Wang, Yizhuo Li, Wenhai Wang, Ping Luo, Yali Wang, Limin Wang, and Yu Qiao. Videochat: Chat-centric video understanding, 2024

    2024

  19. [27]

    Mvbench: A comprehensive multi-modal video understanding benchmark, 2023

    Kunchang Li, Yali Wang, Yinan He, Yizhuo Li, Yi Wang, Yi Liu, Zun Wang, Jilan Xu, Guo Chen, Ping Luo, Limin Wang, and Yu Qiao. Mvbench: A comprehensive multi-modal video understanding benchmark, 2023

    2023

  20. [28]

    Videochat-r1: Enhancing spatio-temporal perception via reinforcement fine-tuning, 2025

    Xinhao Li, Ziang Yan, Desen Meng, Lu Dong, Xiangyu Zeng, Yinan He, Yali Wang, Yu Qiao, Yi Wang, and Limin Wang. Videochat-r1: Enhancing spatio-temporal perception via reinforcement fine-tuning, 2025

    2025

  21. [29]

    Lmeye: An interactive perception network for large language models

    Yunxin Li, Baotian Hu, Xinyu Chen, Lin Ma, Yong Xu, and Min Zhang. Lmeye: An interactive perception network for large language models. IEEE Transactions on Multimedia, 26:10952– 10964, 2024

    2024

  22. [30]

    Uni-moe: Scaling unified multimodal llms with mixture of experts

    Yunxin Li, Shenyuan Jiang, Baotian Hu, Longyue Wang, Wanqi Zhong, Wenhan Luo, Lin Ma, and Min Zhang. Uni-moe: Scaling unified multimodal llms with mixture of experts. IEEE Transactions on Pattern Analysis and Machine Intelligence, 47(5):3424–3439, 2025

    2025

  23. [31]

    Perception, reason, think, and plan: A survey on large multimodal reasoning models

    Yunxin Li, Zhenyu Liu, Zitao Li, Xuanyu Zhang, Zhenran Xu, Xinyu Chen, Haoyuan Shi, Shenyuan Jiang, Xintong Wang, Jifang Wang, et al. Perception, reason, think, and plan: A survey on large multimodal reasoning models. arXiv preprint arXiv:2505.04921, 2025

    2025 arXiv

  24. [32]

    Sti- bench: Are mllms ready for precise spatial-temporal world understanding? arXiv preprint arXiv:2503.23765, 2025

    Yun Li, Yiming Zhang, Tao Lin, XiangRui Liu, Wenxiao Cai, Zheng Liu, and Bo Zhao. Sti- bench: Are mllms ready for precise spatial-temporal world understanding? arXiv preprint arXiv:2503.23765, 2025

    2025 arXiv

  25. [33]

    Vila: On pre-training for visual language models, 2024

    Ji Lin, Hongxu Yin, Wei Ping, Yao Lu, Pavlo Molchanov, Andrew Tao, Huizi Mao, Jan Kautz, Mohammad Shoeybi, and Song Han. Vila: On pre-training for visual language models, 2024. 12

    2024

  26. [34]

    Spatialcot: Advancing spatial reasoning through coordinate alignment and chain-of-thought for embodied task planning, 2025

    Yuecheng Liu, Dafeng Chi, Shiguang Wu, Zhanguang Zhang, Yaochen Hu, Lingfeng Zhang, Yingxue Zhang, Shuang Wu, Tongtong Cao, Guowei Huang, Helong Huang, Guangjian Tian, Weichao Qiu, Xingyue Quan, Jianye Hao, and Yuzheng Zhuang. Spatialcot: Advancing spatial reasoning through co...

    2025

  27. [35]

    Tempcompass: Do video llms really understand videos? arXiv preprint arXiv: 2403.00476, 2024

    Yuanxin Liu, Shicheng Li, Yi Liu, Yuxiang Wang, Shuhuai Ren, Lei Li, Sishuo Chen, Xu Sun, and Lu Hou. Tempcompass: Do video llms really understand videos? arXiv preprint arXiv: 2403.00476, 2024

    2024 arXiv

  28. [36]

    Videomind: A chain-of-lora agent for long video reasoning, 2025

    Ye Liu, Kevin Qinghong Lin, Chang Wen Chen, and Mike Zheng Shou. Videomind: A chain-of-lora agent for long video reasoning, 2025

    2025

  29. [37]

    Seg-zero: Reasoning-chain guided segmentation via cognitive reinforcement, 2025

    Yuqi Liu, Bohao Peng, Zhisheng Zhong, Zihao Yue, Fanbin Lu, Bei Yu, and Jiaya Jia. Seg-zero: Reasoning-chain guided segmentation via cognitive reinforcement, 2025

    2025

  30. [38]

    Video swin transformer, 2021

    Ze Liu, Jia Ning, Yue Cao, Yixuan Wei, Zheng Zhang, Stephen Lin, and Han Hu. Video swin transformer, 2021

    2021

  31. [39]

    Visual-rft: Visual reinforcement fine-tuning, 2025

    Ziyu Liu, Zeyi Sun, Yuhang Zang, Xiaoyi Dong, Yuhang Cao, Haodong Duan, Dahua Lin, and Jiaqi Wang. Visual-rft: Visual reinforcement fine-tuning, 2025

    2025

  32. [40]

    Othink- mr1: Stimulating multimodal generalized reasoning capabilities via dynamic reinforcement learning, 2025

    Zhiyuan Liu, Yuting Zhang, Feng Liu, Changwang Zhang, Ying Sun, and Jun Wang. Othink- mr1: Stimulating multimodal generalized reasoning capabilities via dynamic reinforcement learning, 2025

    2025

  33. [41]

    Gui-r1 : A generalist r1-style vision-language action model for gui agents, 2025

    Run Luo, Lu Wang, Wanwei He, and Xiaobo Xia. Gui-r1 : A generalist r1-style vision-language action model for gui agents, 2025

    2025

  34. [42]

    Video-chatgpt: Towards detailed video understanding via large vision and language models, 2024

    Muhammad Maaz, Hanoona Rasheed, Salman Khan, and Fahad Shahbaz Khan. Video-chatgpt: Towards detailed video understanding via large vision and language models, 2024

    2024

  35. [43]

    Mm-eureka: Exploring the frontiers of multimodal reasoning with rule-based reinforcement learning, 2025

    Fanqing Meng, Lingxiao Du, Zongkai Liu, Zhixiang Zhou, Quanfeng Lu, Daocheng Fu, Tiancheng Han, Botian Shi, Wenhai Wang, Junjun He, Kaipeng Zhang, Ping Luo, Yu Qiao, Qiaosheng Zhang, and Wenqi Shao. Mm-eureka: Exploring the frontiers of multimodal reasoning with rule-based rei...

    2025

  36. [44]

    Video transformer network, 2021

    Daniel Neimark, Omri Bar, Maya Zohar, and Dotan Asselmann. Video transformer network, 2021

    2021

  37. [45]

    Dinov2: Learning robust visual features without supervision, 2024

    Maxime Oquab, Timothée Darcet, Théo Moutakanni, Huy V o, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, Mahmoud Assran, Nicolas Ballas, Wojciech Galuba, Russell Howes, Po-Yao Huang, Shang-Wen Li, Ishan Misra, Michael Rabb...

    2024

  38. [46]

    Spatial-r1: Enhancing mllms in video spatial reasoning, 2025

    Kun Ouyang. Spatial-r1: Enhancing mllms in video spatial reasoning, 2025

    2025

  39. [47]

    Richard Yuanzhe Pang, Alicia Parrish, Nitish Joshi, Nikita Nangia, Jason Phang, Angelica Chen, Vishakh Padmakumar, Johnny Ma, Jana Thompson, He He, and Samuel Bowman. QuALITY: Question answering with long input texts, yes! In Proceedings of the 2022 Conference of the North Ame...

    2022

  40. [48]

    Lmm-r1: Empowering 3b lmms with strong reasoning abilities through two-stage rule-based rl, 2025

    Yingzhe Peng, Gongrui Zhang, Miaosen Zhang, Zhiyuan You, Jie Liu, Qipeng Zhu, Kai Yang, Xingzhong Xu, Xin Geng, and Xu Yang. Lmm-r1: Empowering 3b lmms with strong reasoning abilities through two-stage rule-based rl, 2025

    2025

  41. [49]

    Learning transferable visual models from natural language supervision, 2021

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agar- wal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, and Ilya Sutskever. Learning transferable visual models from natural language supervision, 2021

    2021

  42. [50]

    Manning, and Chelsea Finn

    Rafael Rafailov, Archit Sharma, Eric Mitchell, Stefano Ermon, Christopher D. Manning, and Chelsea Finn. Direct preference optimization: Your language model is secretly a reward model, 2024

    2024

  43. [51]

    Plummer, Ranjay Krishna, Kuo-Hao Zeng, and Kate Saenko

    Arijit Ray, Jiafei Duan, Ellis Brown, Reuben Tan, Dina Bashkirova, Rose Hendrix, Kiana Ehsani, Aniruddha Kembhavi, Bryan A. Plummer, Ranjay Krishna, Kuo-Hao Zeng, and Kate Saenko. Sat: Dynamic spatial aptitude training for multimodal language models, 2025

    2025

  44. [52]

    Proximal policy optimization algorithms, 2017

    John Schulman, Filip Wolski, Prafulla Dhariwal, Alec Radford, and Oleg Klimov. Proximal policy optimization algorithms, 2017

    2017

  45. [53]

    Tomato: Assessing visual temporal reasoning capabilities in multimodal foundation models, 2024

    Ziyao Shangguan, Chuhan Li, Yuxuan Ding, Yanan Zheng, Yilun Zhao, Tesca Fitzgerald, and Arman Cohan. Tomato: Assessing visual temporal reasoning capabilities in multimodal foundation models, 2024. 13

    2024

  46. [54]

    Zhihong Shao, Peiyi Wang, Qihao Zhu, Runxin Xu, Junxiao Song, Xiao Bi, Haowei Zhang, Mingchuan Zhang, Y . K. Li, Y . Wu, and Daya Guo. Deepseekmath: Pushing the limits of mathematical reasoning in open language models, 2024

    2024

  47. [55]

    Efficient reinforcement finetuning via adaptive curriculum learning, 2025

    Taiwei Shi, Yiyang Wu, Linxin Song, Tianyi Zhou, and Jieyu Zhao. Efficient reinforcement finetuning via adaptive curriculum learning, 2025

    2025

  48. [56]

    Mm-verify: Enhancing multimodal reasoning with chain-of-thought verification, 2025

    Linzhuang Sun, Hao Liang, Jingxuan Wei, Bihui Yu, Tianpeng Li, Fan Yang, Zenan Zhou, and Wentao Zhang. Mm-verify: Enhancing multimodal reasoning with chain-of-thought verification, 2025

    2025

  49. [57]

    Reason-rft: Reinforcement fine-tuning for visual reasoning, 2025

    Huajie Tan, Yuheng Ji, Xiaoshuai Hao, Minglan Lin, Pengwei Wang, Zhongyuan Wang, and Shanghang Zhang. Reason-rft: Reinforcement fine-tuning for visual reasoning, 2025

    2025

  50. [58]

    Game-theoretic regularized self-play alignment of large language models, 2025

    Xiaohang Tang, Sangwoong Yoon, Seongho Son, Huizhuo Yuan, Quanquan Gu, and Ilija Bogunovic. Game-theoretic regularized self-play alignment of large language models, 2025

    2025

  51. [59]

    Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context, 2024

    Gemini Team, Petko Georgiev, Ving Ian Lei, Ryan Burnell, et al. Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context, 2024

    2024

  52. [60]

    Gemma 3 technical report, 2025

    Gemma Team, Aishwarya Kamath, Johan Ferret, Shreya Pathak, et al. Gemma 3 technical report, 2025

    2025

  53. [61]

    Kimi-vl technical report, 2025

    Kimi Team, Angang Du, Bohong Yin, Bowei Xing, et al. Kimi-vl technical report, 2025

    2025

  54. [62]

    Model cards & prompt formats-llama 3.2, 2024

    Meta Team. Model cards & prompt formats-llama 3.2, 2024

    2024

  55. [63]

    Vila: On pre-training for visual language models, 2024

    OpenGVLab Team. Vila: On pre-training for visual language models, 2024

    2024

  56. [64]

    Goucher, et al

    OpenAI Team, Aaron Hurst, Adam Lerer, Adam P. Goucher, et al. Gpt-4o system card, 2024

    2024

  57. [65]

    Qwen3: Think deeper, act faster, April 2025

    Qwen Team. Qwen3: Think deeper, act faster, April 2025

    2025

  58. [66]

    Vl- rethinker: Incentivizing self-reflection of vision-language models with reinforcement learning

    Haozhe Wang, Chao Qu, Zuming Huang, Wei Chu, Fangzhen Lin, and Wenhu Chen. Vl- rethinker: Incentivizing self-reflection of vision-language models with reinforcement learning. arXiv preprint arXiv:2504.08837, 2025

    2025 arXiv

  59. [67]

    Piecing it all together: Verifying multi-hop multimodal claims, 2024

    Haoran Wang, Aman Rangapur, Xiongxiao Xu, Yueqing Liang, Haroon Gharwi, Carl Yang, and Kai Shu. Piecing it all together: Verifying multi-hop multimodal claims, 2024

    2024

  60. [68]

    Lvbench: An extreme long video understanding benchmark, 2024

    Weihan Wang, Zehai He, Wenyi Hong, Yean Cheng, Xiaohan Zhang, Ji Qi, Shiyu Huang, Bin Xu, Yuxiao Dong, Ming Ding, and Jie Tang. Lvbench: An extreme long video understanding benchmark, 2024

    2024

  61. [69]

    Sota with less: Mcts-guided sample selection for data-efficient visual reasoning self-improvement, 2025

    Xiyao Wang, Zhengyuan Yang, Chao Feng, Hongjin Lu, Linjie Li, Chung-Ching Lin, Kevin Lin, Furong Huang, and Lijuan Wang. Sota with less: Mcts-guided sample selection for data-efficient visual reasoning self-improvement, 2025

    2025

  62. [70]

    Internvideo2.5: Empowering video mllms with long and rich context modeling, 2025

    Yi Wang, Xinhao Li, Ziang Yan, Yinan He, Jiashuo Yu, Xiangyu Zeng, Chenting Wang, Changlian Ma, Haian Huang, Jianfei Gao, Min Dou, Kai Chen, Wenhai Wang, Yu Qiao, Yali Wang, and Limin Wang. Internvideo2.5: Empowering video mllms with long and rich context modeling, 2025

    2025

  63. [71]

    Unified multimodal chain-of-thought reward model through reinforcement fine-tuning, 2025

    Yibin Wang, Zhimin Li, Yuhang Zang, Chunyu Wang, Qinglin Lu, Cheng Jin, and Jiaqi Wang. Unified multimodal chain-of-thought reward model through reinforcement fine-tuning, 2025

    2025

  64. [72]

    Chain-of-thought prompting elicits reasoning in large language models, 2023

    Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Brian Ichter, Fei Xia, Ed Chi, Quoc Le, and Denny Zhou. Chain-of-thought prompting elicits reasoning in large language models, 2023

    2023

  65. [73]

    Videorope: What makes for good video rotary position embedding?, 2025

    Xilin Wei, Xiaoran Liu, Yuhang Zang, Xiaoyi Dong, Pan Zhang, Yuhang Cao, Jian Tong, Haodong Duan, Qipeng Guo, Jiaqi Wang, Xipeng Qiu, and Dahua Lin. Videorope: What makes for good video rotary position embedding?, 2025

    2025

  66. [74]

    Longvideobench: A benchmark for long-context interleaved video-language understanding, 2024

    Haoning Wu, Dongxu Li, Bei Chen, and Junnan Li. Longvideobench: A benchmark for long-context interleaved video-language understanding, 2024

    2024

  67. [75]

    St-think: How multimodal large language models reason about 4d worlds from ego-centric videos, 2025

    Peiran Wu, Yunze Liu, Miao Liu, and Junxiao Shen. St-think: How multimodal large language models reason about 4d worlds from ego-centric videos, 2025

    2025

  68. [76]

    Self-play preference optimization for language model alignment, 2024

    Yue Wu, Zhiqing Sun, Huizhuo Yuan, Kaixuan Ji, Yiming Yang, and Quanquan Gu. Self-play preference optimization for language model alignment, 2024

    2024

  69. [77]

    Deepseek-vl2: Mixture-of-experts vision-language models for advanced multimodal understanding, 2024

    Zhiyu Wu, Xiaokang Chen, Zizheng Pan, Xingchao Liu, Wen Liu, Damai Dai, Huazuo Gao, Yiyang Ma, Chengyue Wu, Bingxuan Wang, Zhenda Xie, Yu Wu, Kai Hu, Jiawei Wang, Yaofeng Sun, Yukun Li, Yishi Piao, Kang Guan, Aixin Liu, Xin Xie, Yuxiang You, Kai Dong, Xingkai Yu, Haowei Zhang,...

    2024

  70. [78]

    Atomthink: A slow thinking framework for multimodal mathematical reasoning, 2024

    Kun Xiang, Zhili Liu, Zihao Jiang, Yunshuang Nie, Runhui Huang, Haoxiang Fan, Hanhui Li, Weiran Huang, Yihan Zeng, Jianhua Han, Lanqing Hong, Hang Xu, and Xiaodan Liang. Atomthink: A slow thinking framework for multimodal mathematical reasoning, 2024. 14

    2024

  71. [79]

    Echoink-r1: Exploring audio-visual reasoning in multimodal llms via reinforcement learning, 2025

    Zhenghao Xing, Xiaowei Hu, Chi-Wing Fu, Wenhai Wang, Jifeng Dai, and Pheng-Ann Heng. Echoink-r1: Exploring audio-visual reasoning in multimodal llms via reinforcement learning, 2025

    2025

  72. [80]

    Redstar: Does scaling long-cot data unlock better slow-reasoning systems?, 2025

    Haotian Xu, Xing Wu, Weinong Wang, Zhongzhi Li, Da Zheng, Boyuan Chen, Yi Hu, Shijia Kang, Jiaming Ji, Yingying Zhang, Zhijiang Guo, Yaodong Yang, Muhan Zhang, and Debing Zhang. Redstar: Does scaling long-cot data unlock better slow-reasoning systems?, 2025

    2025

  73. [81]

    Thinking in Space: How Multimodal Large Language Models See, Remember and Recall Spaces

    Jihan Yang, Shusheng Yang, Anjali Gupta, Rilyn Han, Li Fei-Fei, and Saining Xie. Thinking in Space: How Multimodal Large Language Models See, Remember and Recall Spaces. arXiv preprint arXiv:2412.14171, 2024

    2024 arXiv

  74. [82]

    R1-onevision: Advancing generalized multimodal reasoning through cross-modal formalization, 2025

    Yi Yang, Xiaoxuan He, Hongkun Pan, Xiyan Jiang, Yan Deng, Xingtao Yang, Haoyu Lu, Dacheng Yin, Fengyun Rao, Minfeng Zhu, Bo Zhang, and Wei Chen. R1-onevision: Advancing generalized multimodal reasoning through cross-modal formalization, 2025

    2025

  75. [83]

    mplug-owl3: Towards long image-sequence understanding in multi-modal large language models, 2024

    Jiabo Ye, Haiyang Xu, Haowei Liu, Anwen Hu, Ming Yan, Qi Qian, Ji Zhang, Fei Huang, and Jingren Zhou. mplug-owl3: Towards long image-sequence understanding in multi-modal large language models, 2024

    2024

  76. [84]

    Dapo: An open-source llm reinforcement learning system at scale, 2025

    Qiying Yu, Zheng Zhang, Ruofei Zhu, Yufeng Yuan, Xiaochen Zuo, Yu Yue, Tiantian Fan, Gaohong Liu, Lingjun Liu, Xin Liu, Haibin Lin, Zhiqi Lin, Bole Ma, Guangming Sheng, Yuxuan Tong, Chi Zhang, Mofan Zhang, Wang Zhang, Hang Zhu, Jinhua Zhu, Jiaze Chen, Jiangjie Chen, Chengyi Wa...

    2025

  77. [85]

    Eric Zelikman, Yuhuai Wu, Jesse Mu, and Noah D. Goodman. Star: Bootstrapping reasoning with reasoning, 2022

    2022

  78. [86]

    Videollama 3: Frontier multimodal foundation models for image and video understanding, 2025

    Boqiang Zhang, Kehan Li, Zesen Cheng, Zhiqiang Hu, Yuqian Yuan, Guanzheng Chen, Sicong Leng, Yuming Jiang, Hang Zhang, Xin Li, Peng Jin, Wenqi Zhang, Fan Wang, Lidong Bing, and Deli Zhao. Videollama 3: Frontier multimodal foundation models for image and video understanding, 2025

    2025

  79. [87]

    Video-llama: An instruction-tuned audio-visual language model for video understanding, 2023

    Hang Zhang, Xin Li, and Lidong Bing. Video-llama: An instruction-tuned audio-visual language model for video understanding, 2023

    2023

  80. [88]

    From flatland to space: Teaching vision-language models to perceive and reason in 3d

    Jiahui Zhang, Yurui Chen, Yanpeng Zhou, Yueming Xu, Ze Huang, Jilin Mei, Junhui Chen, Yujie Yuan, Xinyue Cai, Guowei Huang, Xingyue Quan, Hang Xu, and Li Zhang. From flatland to space: Teaching vision-language models to perceive and reason in 3d. arXiv preprint arXiv:2503.22976, 2025

    2025

  81. [89]

    Long context transfer from language to vision, 2024

    Peiyuan Zhang, Kaichen Zhang, Bo Li, Guangtao Zeng, Jingkang Yang, Yuanhan Zhang, Ziyue Wang, Haoran Tan, Chunyuan Li, and Ziwei Liu. Long context transfer from language to vision, 2024

    2024

  82. [90]

    Tinyllava-video-r1: Towards smaller lmms for video reasoning, 2025

    Xingjian Zhang, Siwei Wen, Wenjun Wu, and Lei Huang. Tinyllava-video-r1: Towards smaller lmms for video reasoning, 2025

    2025

  83. [91]

    Video instruction tuning with synthetic data, 2024

    Yuanhan Zhang, Jinming Wu, Wei Li, Bo Li, Zejun Ma, Ziwei Liu, and Chunyuan Li. Video instruction tuning with synthetic data, 2024

    2024

  84. [92]

    Openrft: Adapting reasoning foundation model for domain-specific tasks with reinforcement fine-tuning, 2024

    Yuxiang Zhang, Yuqi Yang, Jiangming Shu, Yuhang Wang, Jinlin Xiao, and Jitao Sang. Openrft: Adapting reasoning foundation model for domain-specific tasks with reinforcement fine-tuning, 2024

    2024

  85. [93]

    Mme-realworld: Could your multimodal llm challenge high-resolution real-world scenarios that are difficult for humans? arXiv preprint arXiv:2408.13257, 2024

    Yi-Fan Zhang, Huanyu Zhang, Haochen Tian, Chaoyou Fu, Shuangqing Zhang, Junfei Wu, Feng Li, Kun Wang, Qingsong Wen, Zhang Zhang, et al. Mme-realworld: Could your multimodal llm challenge high-resolution real-world scenarios that are difficult for humans? arXiv preprint arXiv:2...

    2024 arXiv

  86. [94]

    Multi- modal chain-of-thought reasoning in language models, 2024

    Zhuosheng Zhang, Aston Zhang, Mu Li, Hai Zhao, George Karypis, and Alex Smola. Multi- modal chain-of-thought reasoning in language models, 2024

    2024

  87. [95]

    R1-omni: Explainable omni-multimodal emotion recognition with reinforcement learning, 2025

    Jiaxing Zhao, Xihan Wei, and Liefeng Bo. R1-omni: Explainable omni-multimodal emotion recognition with reinforcement learning, 2025

    2025

  88. [96]

    Mmvu: Measuring expert-level multi-discipline video understanding, 2025

    Yilun Zhao, Lujing Xie, Haowei Zhang, Guo Gan, Yitao Long, Zhiyuan Hu, Tongyan Hu, Weiyuan Chen, Chuhan Li, Junyang Song, Zhijian Xu, Chengye Wang, Weifeng Pan, Ziyao Shangguan, Xiangru Tang, Zhenwen Liang, Yixin Liu, Chen Zhao, and Arman Cohan. Mmvu: Measuring expert-level mu...

    2025

  89. [97]

    Villa: Video reasoning segmentation with large language model, 2025

    Rongkun Zheng, Lu Qi, Xi Chen, Yi Wang, Kun Wang, Yu Qiao, and Hengshuang Zhao. Villa: Video reasoning segmentation with large language model, 2025. 15

    2025

  90. [98]

    Mlvu: A comprehensive benchmark for multi-task long video understanding

    Junjie Zhou, Yan Shu, Bo Zhao, Boya Wu, Shitao Xiao, Xi Yang, Yongping Xiong, Bo Zhang, Tiejun Huang, and Zheng Liu. Mlvu: A comprehensive benchmark for multi-task long video understanding. arXiv preprint arXiv:2406.04264, 2024. A Detailed Training and Evaluation Analysis A.1 ...

    2024 arXiv

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.