{
 "body": "\n## What it measures\n\nK-MetBench is grounded in national qualification exams and diagnoses chart visual reasoning, expert-verified logical rationales, Korean geo-cultural comprehension and fine-grained meteorological analysis.\n\n## How it is scored\n\nThe abstract reports evaluation of 55 models and identifies modality and reasoning gaps, including stronger performance by Korean models in local contexts.\n\n## Dataset and licence\n\nThe primary source establishes the benchmark release, but the opened materials do not establish a single dataset licence. Confirm the current release terms and split accounting before redistribution.\n\n## Who publishes it\n\nDr. Bench is introduced in the 2026 paper; the opened abstract does not provide a complete author list.\n\n## Lineage\n\nNo predecessor or successor was established in the opened primary source.\n\n## Saturation and contamination\n\nThe benchmark materials are public, which creates contamination opportunities. The opened source does not establish a contamination audit or current saturation ceiling.\n\n## How to run it\n\nFollow the primary paper or release protocol, recording the exact model, prompts, evaluator, task version, tool access and timeout. Preserve per-task outcomes when comparing runs.\n\n## Reading the numbers\n\nHigher scores indicate more successful tasks under the selected protocol. Results can depend on evaluator models, prompts, environment setup and aggregation, so compare only matched configurations.\n\n## Protocol cautions\n\nKeep the task set, context, instructions and evaluator fixed. Report domain-level and difficulty-level results whenever the release defines them, because aggregates can hide systematic failures.\n\nThe benchmark should be treated as a protocol rather than a single model score. Store the exact release revision, context or repository snapshot, prompt, tool permissions, timeout, evaluator version and task-level outcomes. If a task depends on external services or packages, record those dependencies and distinguish an infrastructure failure from an agent failure. Aggregate scores are useful for a headline comparison, while per-domain, per-difficulty and per-task results reveal which capabilities remain unreliable. Public examples and reference material also create opportunities for contamination, so report whether the evaluated model or agent had access to the benchmark before the run.\nThese records therefore leave unspecified details explicitly unknown until the release documentation provides them.\nFuture updates may refine these fields.\n",
 "build": {
  "built_at": "2026-09-09T16:56:50+00:00",
  "commit": "0a599558854c0e238c03a0f0d725239cb28f9d11",
  "eligibility_as_of": "2026-09-09"
 },
 "disposition": {
  "canonical_id": "k_metbench",
  "reasons": [],
  "status": "unassessed",
  "verified_results": []
 },
 "models_covered": [],
 "page": {
  "aliases": [],
  "category": "domain",
  "contamination": {
   "note": "The benchmark materials are public; the opened source does not establish a contamination audit or private rotating holdout.",
   "risk": "medium"
  },
  "dataset": {
   "languages": [
    "en"
   ],
   "license": "",
   "modalities": [
    "text"
   ],
   "public_test_set": true,
   "size": 0,
   "size_note": "The paper reports 0 expert-curated tasks across 10 broad domains.",
   "splits": "Unknown unless specified by the official release.",
   "url": "https://arxiv.org/abs/2604.24645v1"
  },
  "freshness": {
   "researched": "2026-09-09",
   "researched_by": "GPT-5.6 Luna, luna-stream-c-007 (Codex coordinated)",
   "reviewed": "",
   "reviewed_by": ""
  },
  "harness": {
   "bigbench": "",
   "helm": "",
   "inspect_evals": "",
   "lm_eval": "",
   "opencompass": "",
   "other": "Use the primary release protocol and record evaluator settings."
  },
  "id": "k_metbench",
  "last_updated": "",
  "leaderboard_url": "",
  "lineage": {
   "family": "Korean meteorological expertise",
   "predecessor": "",
   "successors": [],
   "variants": []
  },
  "measures": "Expert-level Korean weather forecasting capability across charts, rationales, local context and fine-grained domain analysis.",
  "metric": {
   "baseline_note": "No universal random or human baseline was established in the opened primary source.",
   "direction": "higher_is_better",
   "human_baseline": null,
   "max_score": 100,
   "name": "task success rate",
   "random_baseline": null,
   "unit": "%"
  },
  "name": "K-MetBench",
  "page_kind": "benchmark",
  "paper": {
   "arxiv": "2604.24645",
   "title": "K-MetBench",
   "url": "https://arxiv.org/abs/2604.24645v1",
   "year": 2026
  },
  "publisher": {
   "authors": [],
   "org": "Dr. Bench authors",
   "url": "https://arxiv.org/abs/2604.24645v1"
  },
  "released": "2026-10",
  "repo_url": "",
  "saturation": {
   "as_of": "",
   "note": "No current saturation ceiling was established in the opened source.",
   "status": "unknown",
   "top_score": null
  },
  "sources": [
   {
    "accessed": "2026-09-09",
    "title": "K-MetBench dataset",
    "url": "https://huggingface.co/datasets/soyeonbot/K-MetBench"
   }
  ],
  "status": "active",
  "subcategory": "Korean meteorological expert evaluation",
  "summary": "K-MetBench evaluates multimodal language models for Korean meteorological expertise across visual reasoning, logic, geo-cultural comprehension and domain analysis.",
  "tags": [
   "korean",
   "weather",
   "multimodal",
   "domain"
  ],
  "task_format": "Expert diagnostic questions grounded in national qualification exams, including chart interpretation and Korean domain reasoning."
 }
}