Beacon: Knowing When and How to Perform Agentic Visual Reasoning | Paper Citation Record

Observation f56f4ce3-a20a-4f98-992f-401a0d643fc1 · outbound

This paper cites Qwen3-VL Technical Report.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Qwen3-VL Technical Report

Reference 1

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.572906Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3ad2475c-6393-4ffc-b735-2c1a2b186bf1 · outbound

This paper cites Qwen2.5-VL Technical Report.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Qwen2.5-VL Technical Report

Reference 2

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.577184Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 52f1f3d1-c9a0-418c-9ea2-43d2dd02a6fd · outbound

This paper cites BabyVision: Visual Reasoning Beyond Language.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning BabyVision: Visual Reasoning Beyond Language

Reference 3

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.580613Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 47c4a280-2b81-4b6f-8587-5ab501711019 · outbound

This paper cites Vlmevalkit: An open-source toolkit for evaluating large multi-modality models.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Vlmevalkit: An open-source toolkit for evaluating large multi-modality models

Reference 4

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.583636Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2c34196c-47d9-4041-96d8-3b9f04b581de · outbound

This paper cites Blink: Multimodal large language models can see but not perceive.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Blink: Multimodal large language models can see but not perceive

Reference 5

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.586821Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 03889ca6-0fa5-41e9-94b9-85a991881575 · outbound

This paper cites Gemini 3 flash.https://deepmind.google/models/ model-cards/gemini-3-flash/, 2025.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Gemini 3 flash.https://deepmind.google/models/ model-cards/gemini-3-flash/, 2025

Reference 6

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.589693Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 49c5fe79-91d0-413c-b79d-f1eb21506a92 · outbound

This paper cites Gemini 3.1 Pro: Model Card.https://deepmind.google/ models/model-cards/gemini-3-1-pro/, 2026.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Gemini 3.1 Pro: Model Card.https://deepmind.google/ models/model-cards/gemini-3-1-pro/, 2026

Reference 7

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.592635Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 63ebea04-eff5-4754-93fe-408827e5e25a · outbound

This paper cites Do Multimodal Agents Really Benefit from Tool Use? A Systematic Study of Capability Gains.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Do Multimodal Agents Really Benefit from Tool Use? A Systematic Study of Capability Gains

Reference 8

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.595072Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f5b2e615-6c47-4719-93ae-395fbfa041cd · outbound

This paper cites Olympiadbench: A challenging benchmark for promoting agi with olympiad-level bilingual multimodal scientific problems.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Olympiadbench: A challenging benchmark for promoting agi with olympiad-level bilingual multimodal scientific problems

Reference 9

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.597828Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 88f26d40-cb8c-4c6d-acc5-0c71e6756db8 · outbound

This paper cites Deep- eyesv2: Toward agentic multimodal model.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Deep- eyesv2: Toward agentic multimodal model

Reference 10

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.600434Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 94874c9f-8a53-42e8-a242-eaf352d6884a · outbound

This paper cites Codev: Code with images for faithful visual reasoning via tool-aware policy optimization.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Codev: Code with images for faithful visual reasoning via tool-aware policy optimization

Reference 11

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.603049Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bbcf8ea9-ad85-4c66-91cf-62c90d036a40 · outbound

This paper cites Mini-o3: Scaling Up Reasoning Patterns and Interaction Turns for Visual Search.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Mini-o3: Scaling Up Reasoning Patterns and Interaction Turns for Visual Search

Reference 12

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.605521Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 88b48036-eb45-4c7a-81b8-e4065a270d14 · outbound

This paper cites Tir-bench: A comprehensive benchmark for agen- tic thinking-with-images reasoning.arXiv preprint arXiv:2511.01833, 2025.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Tir-bench: A comprehensive benchmark for agen- tic thinking-with-images reasoning.arXiv preprint arXiv:2511.01833, 2025

Reference 13

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.608319Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e158821a-eb11-4498-a520-58b1e8b14fbc · outbound

This paper cites Inter-gps: Interpretable geometry problem solving with formal language and symbolic reasoning.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Inter-gps: Interpretable geometry problem solving with formal language and symbolic reasoning

Reference 14

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.610922Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d449e6fb-18c7-40d8-8994-12bc5ee509f7 · outbound

This paper cites Mathvista: Evaluating mathemati- cal reasoning of foundation models in visual contexts.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Mathvista: Evaluating mathemati- cal reasoning of foundation models in visual contexts

Reference 15

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.613540Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c1b6ff31-51cf-4e57-937a-0faaf919a127 · outbound

This paper cites What does vision tool-use reinforcement learning really learn? disentangling tool-induced and intrinsic effects for crop-and-zoom.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning What does vision tool-use reinforcement learning really learn? disentangling tool-induced and intrinsic effects for crop-and-zoom

Reference 16

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.616171Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bd8e4216-d41f-46f8-9c28-7d228fb8c16b · outbound

This paper cites Chartqapro: A more diverse and challenging benchmark for chart question answering.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Chartqapro: A more diverse and challenging benchmark for chart question answering

Reference 17

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.618769Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 44f73572-76ad-4cac-a237-d43b714ec9a0 · outbound

This paper cites Adaptive Guidance Accelerates Reinforcement Learning of Reasoning Models.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Adaptive Guidance Accelerates Reinforcement Learning of Reasoning Models

Reference 18

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.621215Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c4e80e5c-8a3b-465d-b450-5aeeb78e4064 · outbound

This paper cites Cogcom: A visual language model with chain-of- manipulations reasoning.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Cogcom: A visual language model with chain-of- manipulations reasoning

Reference 19

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.624249Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 43799434-24a1-4069-b89a-4a6be6d0f474 · outbound

This paper cites V-thinker: Interactive thinking with images.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning V-thinker: Interactive thinking with images

Reference 20

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.626748Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 953cf440-a4c9-4ddf-982e-25a4f36da276 · outbound

This paper cites Vero: An Open RL Recipe for General Visual Reasoning.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Vero: An Open RL Recipe for General Visual Reasoning

Reference 21

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.629138Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation adbe7472-8132-4f29-96a2-c3a856f99321 · outbound

This paper cites Tarr, Aviral Kumar, and Katerina Fragkiadaki.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Tarr, Aviral Kumar, and Katerina Fragkiadaki

Reference 22

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.631786Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2fcb59dc-9646-4e74-9ae7-6760df43ed63 · outbound

This paper cites DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models

Reference 23

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.634206Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2c6a8b91-b34b-4d55-bc16-fd31a665bf5d · outbound

This paper cites HybridFlow: A Flexible and Efficient RLHF Framework.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning HybridFlow: A Flexible and Efficient RLHF Framework

Reference 24

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.637148Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 527d3cd2-3576-4e45-902a-0f13174ab859 · outbound

This paper cites Mavors: Multi-granularity video repre- sentation for multimodal large language model.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Mavors: Multi-granularity video repre- sentation for multimodal large language model

Reference 25

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.639778Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7b7134e3-50b5-4a0d-b17a-6870c7939b9b · outbound

This paper cites Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism

Reference 26

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.642281Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1a1bd232-68f8-4b03-a55a-33150744a8a7 · outbound

This paper cites Codedance: A dynamic tool-integrated mllm for executable visual reason- ing.arXiv preprint arXiv:2512.17312, 2025.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Codedance: A dynamic tool-integrated mllm for executable visual reason- ing.arXiv preprint arXiv:2512.17312, 2025

Reference 27

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.645248Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6b89ba06-8336-4c64-8b10-f79a53b39f0d · outbound

This paper cites VisualPuzzles: Decoupling Multimodal Reasoning Evaluation from Domain Knowledge.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning VisualPuzzles: Decoupling Multimodal Reasoning Evaluation from Domain Knowledge

Reference 28

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.647722Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a2211762-9c6f-4c68-a745-0747bf52ce8d · outbound

This paper cites Pixel reasoner: In- centivizing pixel space reasoning via curiosity-driven reinforcement learning.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Pixel reasoner: In- centivizing pixel space reasoning via curiosity-driven reinforcement learning

Reference 29

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.650544Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation de4762bb-fa75-4014-a8dd-3c9342da3222 · outbound

This paper cites OpenThinkIMG: Learning to Think with Images via Visual Tool Reinforcement Learning.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning OpenThinkIMG: Learning to Think with Images via Visual Tool Reinforcement Learning

Reference 30

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.652956Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ba5d70e8-f92f-4510-8d44-0c8b79ffe84c · outbound

This paper cites Agentvista: Evaluating multimodal agents in ultra-challenging realistic visual scenarios.arXiv preprint arXiv:2602.23166, 2026.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Agentvista: Evaluating multimodal agents in ultra-challenging realistic visual scenarios.arXiv preprint arXiv:2602.23166, 2026

Reference 31

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.655849Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0ef0366d-2e5c-4234-b987-a5d035300038 · outbound

This paper cites Game-rl: Synthesizing multimodal verifiable game data to boost vlms’ general reasoning.arXiv preprint arXiv:2505.13886, 2025.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Game-rl: Synthesizing multimodal verifiable game data to boost vlms’ general reasoning.arXiv preprint arXiv:2505.13886, 2025

Reference 32

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.658361Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 09e21fdd-2760-4efc-b433-5ebe9d85e668 · outbound

This paper cites Cambrian-1: A fully open, vision-centric exploration of multimodal llms.NeurIPS, 2024.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Cambrian-1: A fully open, vision-centric exploration of multimodal llms.NeurIPS, 2024

Reference 33

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.661005Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 389bd31a-a17d-470f-9d2b-8523ec521cf9 · outbound

This paper cites AdaTooler-V: Adaptive Tool-Use for Images and Videos.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning AdaTooler-V: Adaptive Tool-Use for Images and Videos

Reference 34

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.663533Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 40ac1b5b-ab6a-4a45-9b86-9aff181e1c2a · outbound

This paper cites Muirbench: A comprehensive benchmark for robust multi-image understanding.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Muirbench: A comprehensive benchmark for robust multi-image understanding

Reference 35

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.666643Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a67f7c40-d932-4b55-a677-e2d37e1f26df · outbound

This paper cites Measuring multimodal mathematical reasoning with math-vision dataset.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Measuring multimodal mathematical reasoning with math-vision dataset

Reference 36

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.669221Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8fdb219b-c773-4c1c-a0a3-14b135fc928f · outbound

This paper cites Monet: Reasoning in latent visual space beyond images and language.arXiv preprint arXiv:2511.21395, 2025.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Monet: Reasoning in latent visual space beyond images and language.arXiv preprint arXiv:2511.21395, 2025

Reference 37

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.671771Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2f146334-1353-4426-8eab-29b9b2fffe81 · outbound

This paper cites Divide, conquer and combine: A training-free framework for high-resolution image perception in multimodal large language models.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Divide, conquer and combine: A training-free framework for high-resolution image perception in multimodal large language models

Reference 38

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.674232Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1de094e6-1267-4e90-9c1d-53e7018a05b3 · outbound

This paper cites V*: Guided visual search as a core mechanism in multimodal llms.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning V*: Guided visual search as a core mechanism in multimodal llms

Reference 39

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.676776Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a04476fa-5229-489e-8479-8c7cae3830bb · outbound

This paper cites RealWorldQA.https://huggingface.co/datasets/xai-org/ RealworldQA, 2024.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning RealWorldQA.https://huggingface.co/datasets/xai-org/ RealworldQA, 2024

Reference 40

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.679355Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e21fe376-6fce-48b1-ae9a-58a89fe1128a · outbound

This paper cites Act Wisely: Cultivating Meta-Cognitive Tool Use in Agentic Multimodal Models.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Act Wisely: Cultivating Meta-Cognitive Tool Use in Agentic Multimodal Models

Reference 41

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.681811Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4ea6e06d-a2ab-4349-9a6b-858e3a176dbf · outbound

This paper cites Position: Your VLM may not be thinking with interleaved images.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Position: Your VLM may not be thinking with interleaved images

Reference 42

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.684618Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 287dad20-df4c-4143-8bb4-4f189ae76de4 · outbound

This paper cites Mmmu: A massive multi-discipline multi- modal understanding and reasoning benchmark for expert agi.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Mmmu: A massive multi-discipline multi- modal understanding and reasoning benchmark for expert agi

Reference 43

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.687874Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f741e33d-8ee2-426a-9343-ec3333f347c6 · outbound

This paper cites Thyme: Think beyond images.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Thyme: Think beyond images

Reference 44

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.690467Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6d2aae7e-0ea5-4e7b-be9a-2ae0bb25adaf · outbound

This paper cites Hrscene: How far are vlms from effective high-resolution image understanding? InICCV, 2025.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Hrscene: How far are vlms from effective high-resolution image understanding? InICCV, 2025

Reference 45

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.693046Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cf2cda23-47e6-4e56-a544-3098b1644db9 · outbound

This paper cites Pyvision-rl: Forging open agentic vision models via rl.arxiv preprint arxiv:2602.20739, 2026.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Pyvision-rl: Forging open agentic vision models via rl.arxiv preprint arxiv:2602.20739, 2026

Reference 46

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.695407Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c6c228fe-01d3-4654-97d2-df6d8e3e858f · outbound

This paper cites SWIFT:A Scalable lightWeight Infrastructure for Fine-Tuning.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning SWIFT:A Scalable lightWeight Infrastructure for Fine-Tuning

Reference 47

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.697827Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 35e422da-b073-42c1-973c-dc239c450aad · outbound

This paper cites Chain-of-focus prompting: Leveraging sequential visual cues to prompt large autoregres- sive vision models.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Chain-of-focus prompting: Leveraging sequential visual cues to prompt large autoregres- sive vision models

Reference 48

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.700631Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e5fb761e-292a-4ff6-8258-445359099750 · outbound

This paper cites Vtc-bench: Evaluating agentic multi- modal models via compositional visual tool chaining.arXiv preprint arXiv:2603.15030, 2026.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Vtc-bench: Evaluating agentic multi- modal models via compositional visual tool chaining.arXiv preprint arXiv:2603.15030, 2026

Reference 49

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.703141Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e283a379-a6d1-492f-8588-806fe4e60f67 · outbound

This paper cites Retention.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Retention

Reference 50

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.706071Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8bb8e1ba-b7fc-445b-9999-b2e6d93d00c6 · outbound

This paper cites an unresolved cited work.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Unresolved cited work

Reference 51

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.709130Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f97874b0-1497-4adb-804b-d85d924f96c2 · outbound

This paper cites Original question images use stable ids like q0, q1, and images returned by earlier code calls use stable ids like g0, g1.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Original question images use stable ids like q0, q1, and images returned by earlier code calls use stable ids like g0, g1

Reference 52

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.711746Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e27d8c12-2499-4c76-b4cf-c7e7884f604d · outbound

This paper cites For example: annotated = get_image("q0").copy().

Beacon: Knowing When and How to Perform Agentic Visual Reasoning For example: annotated = get_image("q0").copy()

Reference 53

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.714376Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cee45430-5898-4b3f-958c-d5e737bb7dbd · outbound

This paper cites If the current call should return a numeric result, the last line should be the numeric object itself, such as result.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning If the current call should return a numeric result, the last line should be the numeric object itself, such as result

Reference 54

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.716990Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 21cc2461-cffe-457e-a867-b715eea560b0 · outbound

This paper cites Use Python code.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Use Python code

Reference 55

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.719651Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 73930723-dec4-41b8-a427-c0242503e958 · outbound

This paper cites It is not something you should generate yourself.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning It is not something you should generate yourself

Reference 56

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.722348Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ef9e1152-14a7-48c9-a488-0d8411ff662c · outbound

This paper cites an unresolved cited work.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Unresolved cited work

Reference 57

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.724852Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b6a5a8e6-11ea-43b5-a58b-039cb925a8de · outbound

This paper cites After that, you may continue with additional <tool_call> -> <tool_response> -> <observation> cycles.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning After that, you may continue with additional <tool_call> -> <tool_response> -> <observation> cycles

Reference 58

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.727297Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7815045d-5fb0-4cb1-9d52-576387c63d44 · outbound

This paper cites q0").copy() draw = ImageDraw.Draw(annotated) draw.rectangle((600, 290, 660, 380), outline=.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning q0").copy() draw = ImageDraw.Draw(annotated) draw.rectangle((600, 290, 660, 380), outline=

Reference 59

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.729967Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 58918079-77b7-49ed-ab0a-e69a77af3b4a · outbound

This paper cites Work in progress.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Work in progress

Reference 60

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.732686Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9795a361-567c-4b45-9b9e-15562723c015 · outbound

This paper cites an unresolved cited work.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Unresolved cited work

Reference 61

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.735233Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2140fda0-a99d-4439-8270-067d962d2f2b · outbound

This paper cites an unresolved cited work.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Unresolved cited work

Reference 62

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.737951Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f8aa26af-7a43-48e4-8967-fc9c68b677e7 · outbound

This paper cites an unresolved cited work.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Unresolved cited work

Reference 63

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.740437Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation dc8115cf-ab02-46f3-b588-6613b10fec7b · outbound

This paper cites an unresolved cited work.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Unresolved cited work

Reference 64

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.742918Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 01934e87-6841-4b7e-b782-b5d8748c1a3f · outbound

This paper cites Your task:.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Your task:

Reference 65

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.745498Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 721699a6-5bbe-4c31-be63-af49893c64fe · outbound

This paper cites If the tools did not provide helpful information and the answer was effectively derived from tool-free reasoning, especially in the last step, discard the trajectory.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning If the tools did not provide helpful information and the answer was effectively derived from tool-free reasoning, especially in the last step, discard the trajectory

Reference 66

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.748223Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6481be4f-4d3b-44bd-929a-4bdf040447c7 · outbound

This paper cites Remove the unhelpful repeated tool-call steps and keep only the step that produced the key useful information.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Remove the unhelpful repeated tool-call steps and keep only the step that produced the key useful information

Reference 67

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.751038Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8e132085-1fd6-4639-9a8d-f4b0555c011c · outbound

This paper cites an unresolved cited work.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Unresolved cited work

Reference 68

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.753539Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 594813a7-0842-4cd8-a4c2-804ec02113fe · outbound

This paper cites <tool_call>result = 1result</tool_call>.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning <tool_call>result = 1result</tool_call>

Reference 69

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.756021Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2ada38ae-0135-49b7-af68-e6eb62708cb7 · outbound

This paper cites an unresolved cited work.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Unresolved cited work

Reference 70

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.758925Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 23ff0178-155a-4c30-ab85-16d9eba831f2 · outbound

This paper cites an unresolved cited work.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Unresolved cited work

Reference 71

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.761488Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5701dac1-004c-480d-a7bd-253319c6baa1 · outbound

This paper cites If not, try to refine the <observation> to make it more accurate and specific to the actual tool response.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning If not, try to refine the <observation> to make it more accurate and specific to the actual tool response

Reference 72

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.764216Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ba34f0f3-8a81-4021-b255-30062c044d0d · outbound

This paper cites decision.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning decision

Reference 73

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.766799Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4064dee0-ee66-47ed-b795-6d7fc8a69e53 · outbound

This paper cites decision.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning decision

Reference 74

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.769462Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 11db76c3-a617-49ac-819a-608158d109c7 · outbound

This paper cites decision.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning decision

Reference 75

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.772299Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 387665f0-ab5a-42e6-a3ab-fe82aba7959b · outbound

This paper cites messages.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning messages

Reference 76

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.774898Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 16458e43-b5c6-429a-a411-35af7f8ba609 · outbound

This paper cites kind":"original_message.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning kind":"original_message

Reference 77

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.777672Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b5dcd42d-12af-4582-a4db-31ec8f260ed6 · outbound

This paper cites If a tool- response step is kept, keep its <tool_response> in the corresponding original_message instead of rewriting it.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning If a tool- response step is kept, keep its <tool_response> in the corresponding original_message instead of rewriting it

Reference 78

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.780502Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4e132a52-4ee8-45e3-8585-fb058a72411e · outbound

This paper cites an unresolved cited work.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Unresolved cited work

Reference 79

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.783235Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1c756479-ba4d-4509-a6e2-f77f415d3881 · outbound

This paper cites You should decide whether to use code.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning You should decide whether to use code

Reference 80

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.785718Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 240dd7ca-3c59-4002-a424-8e8e5317b7e9 · outbound

This paper cites Only write codes when necessary.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Only write codes when necessary

Reference 81

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.788386Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 97dd0d38-a0d2-4bd4-98a0-77ee01622d77 · outbound

This paper cites hint"field as the training hint. The remaining fields,.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning hint"field as the training hint. The remaining fields,

Reference 82

Resolution: malformed identifier; no resolver link, observed 2026-07-31T02:45:28.790836Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4886ad11-f260-487d-b221-8e0ac8861091 · outbound

This paper cites an unresolved cited work.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Unresolved cited work

Reference 83

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.793791Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 715a1468-ea82-4820-abb6-38653cf7895b · outbound

This paper cites q0") ... <tool call> from PIL import ImageDraw annotated = get image(.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning q0") ... <tool call> from PIL import ImageDraw annotated = get image(

Reference 84

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.796309Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0afa8493-ba4e-40b2-b8a4-7a72e6ce0d4a · outbound

This paper cites Original image refs: q0 Generated image refs: g0 </tool response> <observation> The resized image clearly shows the text "CVPR" in the center of the spiral pattern.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Original image refs: q0 Generated image refs: g0 </tool response> <observation> The resized image clearly shows the text "CVPR" in the center of the spiral pattern

Reference 85

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.798852Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 27baaa7b-5ede-46d3-93f1-6d18a5738f01 · outbound

This paper cites an unresolved cited work.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Unresolved cited work

Reference 86

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.801757Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a1b47cb7-aeff-4312-8b3a-4fa4ddbf2325 · outbound

This paper cites an unresolved cited work.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning Unresolved cited work

Reference 87

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.804274Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2c5fcaf0-3772-4c99-9512-25faf8bf9fab · outbound

This paper cites What appears in the field with the question mark if we do the same with the number 5?.

Beacon: Knowing When and How to Perform Agentic Visual Reasoning What appears in the field with the question mark if we do the same with the number 5?

Reference 88

Resolution: unresolved; no resolver link, observed 2026-07-31T02:45:28.806859Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Coverage vector

Reference resolution

Outbound references

Pith citing papers