{ "schema_version": "paper_public_manifest_v1", "paper_id": "archon-whole-body-intelligence_2026_07", "slug": "archon-whole-body-intelligence", "title": "Teacher Li Hongyang's Whole-Body Intelligence Surpasses GR's Whole-Body Control", "authors": [], "source": { "arxiv_id": "", "pdf_url": "", "project_url": "https://archon.tech/blog/whole-body-intelligence-cn", "github_url": "", "huggingface_url": "", "original_source": "https://archon.tech/blog/whole-body-intelligence-cn" }, "site": { "post_url": "/posts/archon-whole-body-intelligence", "canonical_url": "https://haiguangboy.com/posts/archon-whole-body-intelligence", "cover_image": "https://static.haiguangboy.com/papers/archon-whole-body-intelligence/cover.webp" }, "taxonomy": { "domain": "embodied_ai", "track": "vla", "tasks": [ "embodied_ai", "vla", "action_generation", "robotics", "state_prediction", "Embodied Intelligence", "Humanoid Robots", "Whole-Body Intelligence" ], "related_topics": [ { "paper_id": "an_open_foundation_model_towards_2026_07", "title": "An_Open_Foundation_Model_Towards", "url": "https://haiguangboy.com/posts/an_open_foundation_model_towards", "relation": "contrast", "summary": "Definition: Whole-body intelligence is a pretraining paradigm, explicitly not equivalent to whole-body control, nor is it assembling VLA/BFM/controllers together; contradicts joint co-training of heterogeneous human-robot data as a structurally suboptimal approach", "strength": "strong" }, { "paper_id": "latepost_xuhuazhe_202603_2026_03", "title": "latepost_xuhuazhe_202603", "url": "https://haiguangboy.com/posts/latepost_xuhuazhe_202603", "relation": "contrast", "summary": "Four-layer stack S2/S1/S0.5/S0, each layer with clear interfaces; contradicts route bet: behavior/action parts must be a unified model, opposing modular assembly", "strength": "strong" }, { "paper_id": "wx_界面新闻_20260605_2026_06", "title": "wx_Interface News_20260605", "url": "https://haiguangboy.com/posts/wx_界面新闻_20260605", "relation": "contrast", "summary": "Middle-ground stance on hardware-software coupling: hardware-aware pretraining, neither pursuing hardware independence nor allowing unchecked coupling; contradicts route bet: building the brain, not the body—betting on the foundation model company identity, rejecting the athletic capability track", "strength": "strong" }, { "paper_id": "deepmind_blog_gemini_robotics_2_brings_20260730_2026_07", "title": "The Overlooked Non-Consensus Is Unified Whole-Body Control", "url": "https://haiguangboy.com/posts/gemini-robotics-2", "relation": "contrast", "summary": "S1's key claim: native humanoid is not a robotic arm plus a mobile base, nor a VLM followed by an action head; contradicts architecture: releasing three models at once, layered rather than end-to-end integrated", "strength": "strong" }, { "paper_id": "gemini_robotics_2_safety_2026_07", "title": "Gemini_Robotics_2_Safety", "url": "https://haiguangboy.com/posts/gemini_robotics_2_safety", "relation": "same_track", "summary": "Division of labor statement: whole-body control solves execution problems, whole-body intelligence solves body organization problems; validates empirical finding: there is a clear gap between models 'understanding constraints' and 'acting according to constraints'", "strength": "strong" }, { "paper_id": "google_blog_gemini_robotics_er_2_20260730_2026_07", "title": "Introducing Gemini Robotics ER 2", "url": "https://haiguangboy.com/posts/google_blog_gemini_robotics_er_2_20260730", "relation": "same_track", "summary": "Four-layer stack S2/S1/S0.5/S0, each layer with clear interfaces; validates architecture statement: VLA model and navigation API are declared as agent tools in parallel", "strength": "strong" } ] }, "analyst_take": { "type": "author_opinion", "text": "The most interesting part is not where it agrees, but that it takes opposing stances on two levels\n\n📚 On the question of 'should it be decomposed,' it aligns with the faction advocating unified models and opposing modular assembly—both believe humanoid manipulation cannot be cut along module boundaries. Yet the four-layer stack it proposes is precisely a layered architecture, aligning with GR's approach of separating upper layers\n\n🧲 These two things are not contradictory because they address different cuts. It opposes cutting functions horizontally like 'navigation module plus grasping module,' and advocates cutting abstraction levels vertically like 'semantics, body priors, motion generation, execution.' The former breaks coupled actions apart; the latter does not\n\n🎯 So this judgment can be more precise: the disagreement is not about whether to layer, but whether the cut is at functional boundaries or abstraction levels. Cutting at functional boundaries assigns walking and reaching to two models; cutting at abstraction levels keeps whole-body coordination within the same layer\n\n🧬 Another point worth noting separately: it argues the model must know what hardware it stands on—joint reachable ranges, force control precision, sensor perspectives must all enter the priors. This is a bet in the opposite direction from 'the same model swaps bodies in hours,' and we will see who is right in six months!" }, "ruling": { "importance_score": 3.0, "one_sentence": "Teacher Li Hongyang's Whole-Body Intelligence Surpasses GR's Whole-Body Control" }, "asset_base_url": "https://static.haiguangboy.com/papers/archon-whole-body-intelligence", "assets": [ { "type": "cover_image", "object_key": "papers/archon-whole-body-intelligence/cover.webp", "content_type": "image/webp", "upload_status": "uploaded", "bucket": "paper-assets", "url": "https://static.haiguangboy.com/papers/archon-whole-body-intelligence/cover.webp", "role": "post_cover", "size_bytes": 63812 }, { "type": "public_brief", "object_key": "papers/archon-whole-body-intelligence/public_brief.md", "content_type": "text/markdown; charset=utf-8", "upload_status": "uploaded", "bucket": "paper-assets", "url": "https://static.haiguangboy.com/papers/archon-whole-body-intelligence/public_brief.md", "role": "public_brief", "size_bytes": 4107 }, { "type": "public_manifest", "object_key": "papers/archon-whole-body-intelligence/public_manifest.json", "content_type": "application/json; charset=utf-8", "upload_status": "uploaded", "bucket": "paper-assets", "url": "https://static.haiguangboy.com/papers/archon-whole-body-intelligence/public_manifest.json", "role": "public_manifest", "size_bytes": 6512 } ], "published_at": "2026-08-02T06:51:40+08:00", "created_at": "2026-08-02T06:51:40+08:00", "updated_at": "2026-09-02T10:55:06+08:00" }