Pith Number

pith:TIR5IYQ4

pith:2025:TIR5IYQ4BT5UU2TYPFXD5LXIS5

not attested not anchored not stored refs resolved

Beyond Classification Accuracy: Neural-MedBench and the Need for Deeper Reasoning Benchmarks

Huan Gao, Junling Lin, Mengting Jia, Miao Jing, Mingkun Xu, Shangyang Li, Zhongxia Shen

Vision-language models show major reasoning shortfalls on a new compact neurology benchmark despite high scores on standard tests.

arxiv:2509.22258 v5 · 2025-09-26 · cs.CV · cs.AI

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{TIR5IYQ4BT5UU2TYPFXD5LXIS5}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more

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

Through systematic evaluation of state-of-the-art VLMs, including GPT-4o, Claude-4, and MedGemma, we observe a sharp performance drop compared to conventional datasets. Error analysis shows that reasoning failures, rather than perceptual errors, dominate model shortcomings.

C2weakest assumption

The hybrid scoring pipeline (LLM-based graders combined with clinician validation and semantic similarity metrics) provides a reliable and unbiased measure of true clinical reasoning ability rather than introducing grader-specific artifacts or inconsistencies.

C3one line summary

Neural-MedBench reveals sharp performance drops in state-of-the-art VLMs on reasoning-intensive neurology tasks compared to conventional classification benchmarks, with reasoning failures dominating errors.

References

42 extracted · 42 resolved · 7 Pith anchors

[1] Anthropic. Claude haiku, 2024. URL https://www.anthropic.com 2024

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

[3] The liver tumor segmentation benchmark (lits) 1901

[4] Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities 2025 · arXiv:2507.06261

[5] Benchmarking generative ai for scoring medical student interviews in objective structured clinical examinations (osces) 2025

Formal links

2 machine-checked theorem links

Receipt and verification

First computed	2026-05-20T01:04:58.473052Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

9a23d4621c0cfb4a6a78796e3eaee8974ba6f07fa10719a963535b3d4de74b80

Aliases

arxiv: 2509.22258 · arxiv_version: 2509.22258v5 · doi: 10.48550/arxiv.2509.22258 · pith_short_12: TIR5IYQ4BT5U · pith_short_16: TIR5IYQ4BT5UU2TY · pith_short_8: TIR5IYQ4

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/TIR5IYQ4BT5UU2TYPFXD5LXIS5 \
  | 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: 9a23d4621c0cfb4a6a78796e3eaee8974ba6f07fa10719a963535b3d4de74b80

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "078d657566ca09838af6eaf4c59464c96e429a6955d5c2e011ef74d1978f4ac1",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.CV",
    "submitted_at": "2025-09-26T12:20:01Z",
    "title_canon_sha256": "166b1aa83792ea3889f8f6c7ccba8e1572dc2a28c0c56ea32c0dd9d82e97401f"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2509.22258",
    "kind": "arxiv",
    "version": 5
  }
}