Pith Number

pith:TQMC7C3Z

pith:2025:TQMC7C3ZTOH53UI6BQLC525C7S

not attested not anchored not stored refs resolved

IPR-1: Interactive Physical Reasoner

Guocan Xie, Jiting Cai, Lifeng Zhuo, Mingyu Zhang, Renjie Zhao, Tianxi Tan, Xian Nie, Yan Li, Yong-Lu Li, Ziyu Wang, Zizhu He

An interactive physical reasoner learns causal physics from game play and surpasses GPT-5 overall.

arxiv:2511.15407 v3 · 2025-11-19 · cs.AI · cs.CV · cs.LG

Open paper page JSON Open Graph Bundle Merged state What is a Pith Number?

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

Pretrained on 1,000+ games, our IPR performs robustly on levels from primitive intuition to goal-driven reasoning, and even surpasses GPT-5 overall. We find that performance improves with more training games and interaction steps, and that the model also zero-shot transfers to unseen games.

C2weakest assumption

That world-model rollouts capture true underlying physics and causality rather than visual patterns, and that the G2U benchmark's visual domain gaps and heterogeneous games sufficiently isolate core mechanisms from superficial appearance.

C3one line summary

IPR uses world-model rollouts to reinforce a VLM policy via PhysCode on a 1000+ game benchmark, achieving robust physical reasoning that improves with experience and transfers zero-shot to unseen games while surpassing GPT-5.

References

88 extracted · 88 resolved · 11 Pith anchors

[1] Do as i can, not as i say: Grounding language in robotic affordances, 2022 2022

[2] Metric space magnitude and generalisation in neural networks 2023

[3] V-jepa 2: Self-supervised video models enable understanding, prediction and planning, 2025 2025

[4] V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planning 2025 · arXiv:2506.09985

[5] Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond 2023 · arXiv:2308.12966

Formal links

2 machine-checked theorem links

Cited by

2 papers in Pith

Towards Generalist Game Players: An Investigation of Foundation Models in the Game Multiverse

Receipt and verification

First computed	2026-05-17T23:39:17.119847Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

9c182f8b799b8fddd11e0c162eeba2fc9c684c9ca39d7f4674f6ff3c83525088

Aliases

arxiv: 2511.15407 · arxiv_version: 2511.15407v3 · doi: 10.48550/arxiv.2511.15407 · pith_short_12: TQMC7C3ZTOH5 · pith_short_16: TQMC7C3ZTOH53UI6 · pith_short_8: TQMC7C3Z

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/TQMC7C3ZTOH53UI6BQLC525C7S \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: 9c182f8b799b8fddd11e0c162eeba2fc9c684c9ca39d7f4674f6ff3c83525088

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "e2416c25fff8345fc89728c0fe0b05185a418e8102d0379253dca303700d4fad",
    "cross_cats_sorted": [
      "cs.CV",
      "cs.LG"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.AI",
    "submitted_at": "2025-11-19T13:04:44Z",
    "title_canon_sha256": "db3f1667fba7e9743158fd5fde8fafc3be7a2cb8af83b51d4eb65032bf758297"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2511.15407",
    "kind": "arxiv",
    "version": 3
  }
}