Dataset Viewer
The dataset viewer is not available for this subset.
Cannot get the split names for the config 'default' of the dataset.
Exception:    SplitsNotFoundError
Message:      The split names could not be parsed from the dataset config.
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 290, in _generate_tables
                  pa_table = paj.read_json(
                      io.BytesIO(batch), read_options=paj.ReadOptions(block_size=block_size)
                  )
                File "pyarrow/_json.pyx", line 342, in pyarrow._json.read_json
                File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
                File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
                  raise convert_status(status)
              pyarrow.lib.ArrowInvalid: JSON parse error: Column() changed from object to number in row 0
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 286, in get_dataset_config_info
                  for split_generator in builder._split_generators(
                                         ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      StreamingDownloadManager(base_path=builder.base_path, download_config=download_config)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 101, in _split_generators
                  pa_table = next(iter(self._generate_tables(**splits[0].gen_kwargs, allow_full_read=False)))[1]
                             ~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 304, in _generate_tables
                  batch = json_encode_fields_in_json_lines(original_batch, json_field_paths)
                File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 111, in json_encode_fields_in_json_lines
                  examples = [ujson_loads(line) for line in original_batch.splitlines()]
                              ~~~~~~~~~~~^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 20, in ujson_loads
                  return pd.io.json.ujson_loads(*args, **kwargs)
                         ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
              ValueError: Trailing data
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/split_names.py", line 68, in compute_split_names_from_streaming_response
                  for split in get_dataset_split_names(
                               ~~~~~~~~~~~~~~~~~~~~~~~^
                      path=dataset,
                      ^^^^^^^^^^^^^
                      config_name=config,
                      ^^^^^^^^^^^^^^^^^^^
                      token=hf_token,
                      ^^^^^^^^^^^^^^^
                  )
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 340, in get_dataset_split_names
                  info = get_dataset_config_info(
                      path,
                  ...<6 lines>...
                      **config_kwargs,
                  )
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 291, in get_dataset_config_info
                  raise SplitsNotFoundError("The split names could not be parsed from the dataset config.") from err
              datasets.inspect.SplitsNotFoundError: The split names could not be parsed from the dataset config.

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

The Price of Intelligence

A quality-adjusted price panel for AI inference services (API only), built from public data at zero cost, with the capability table and crosswalk it is joined to, and the validity audit of the benchmarks underneath it.

What is here

File Rows What it is
data/panel/panel.parquet 21,791 Posted prices, one row per model x reasoning effort x provider x month x source
data/panel/benchmark_scores.parquet 5,067 Capability table, 4,605 scored cells on one bounded scale
data/panel/model_crosswalk.csv 1,350 Panel-to-Epoch namespace join, method recorded per row
data/panel/codebook.md Column-by-column codebook

The headline panel is 21,024 observed price observations from 2024-02 to 2026-08, across 3,208 priced models, 86 providers and 31 months. The crosswalk reaches 500 of 782 scored models.

How to read it honestly

  • Prices are posted list prices, on one named channel per provider. Transaction prices sit below list, so the true decline is faster than the panel shows.
  • Gaps are gaps. A month with no observation carries price_flag = gap and a gap_reason. Nothing is interpolated or carried forward.
  • Every observation carries source_url and archive_timestamp, and the provenance ledgers in the Zenodo deposit give one line per observation.
  • Effort variants share a posted price and carry their own capability and tokens per task. One posted price is one observation: weight rows 1/k across the k effort variants of an endpoint and cluster on the endpoint.
  • The capability measure is estimated, not observed. Use the plausible values, not a point estimate, if you want honest intervals.

Provenance and reproduction

The paper this dataset supports is a preprint, arXiv:2608.29843.

Version 1.0.1 regenerates the full analysis chain from the bundle alone; archived at 10.5281/zenodo.22693256.

The analysis was pre-registered before estimation began: https://osf.io/5uqj2/, DOI 10.17605/OSF.IO/5UQJ2. Every deviation from it is logged in the decision log shipped with the deposit.

The git repository omits the 29 provenance ledgers (85 MB); data/raw/MANIFEST.sha256 stands in for them there. The ledgers ship with the Zenodo deposit, and reproduction runs from that bundle, not from a bare clone. Verify with shasum -a 256 -c MANIFEST.sha256.

Licence

Data CC-BY-4.0, code MIT. See LICENSE-DATA.txt and LICENSE-CODE.txt. Attribution string in LICENSE-DATA.txt.

Croissant metadata

{
  "@context": {
    "@vocab": "https://schema.org/",
    "cr": "http://mlcommons.org/croissant/",
    "sc": "https://schema.org/",
    "data": {
      "@id": "cr:data",
      "@type": "@json"
    }
  },
  "@type": "sc:Dataset",
  "conformsTo": "http://mlcommons.org/croissant/1.0",
  "name": "price-of-intelligence",
  "description": "Quality-adjusted price panel for AI inference services, with the capability table and crosswalk it is joined to.",
  "license": "https://creativecommons.org/licenses/by/4.0/",
  "citation": "Zhu, Louis Yiven (2026). The Price of Intelligence. https://doi.org/10.5281/zenodo.22177190",
  "url": "https://huggingface.co/datasets/louisyzhu/price-of-intelligence",
  "version": "1.0.0",
  "distribution": [
    {
      "@type": "cr:FileObject",
      "@id": "panel.parquet",
      "name": "panel.parquet",
      "description": "Price panel: one row per model x reasoning effort x provider x month x source.",
      "contentUrl": "data/panel/panel.parquet",
      "encodingFormat": "application/vnd.apache.parquet",
      "sha256": "23a8c90177c6bcc2af36e24e443a39c73866cfb93a266a9f3f3a9e51e2a8ac05"
    },
    {
      "@type": "cr:FileObject",
      "@id": "benchmark_scores.parquet",
      "name": "benchmark_scores.parquet",
      "description": "Capability table: one row per model x benchmark, on one bounded scale.",
      "contentUrl": "data/panel/benchmark_scores.parquet",
      "encodingFormat": "application/vnd.apache.parquet",
      "sha256": "8f92c5f43214f8eb11b9c0977ca30d26bb05ac2984c9a38fa571c48651a2c9a8"
    },
    {
      "@type": "cr:FileObject",
      "@id": "model_crosswalk.csv",
      "name": "model_crosswalk.csv",
      "description": "Crosswalk between the panel and Epoch namespaces, method recorded per row.",
      "contentUrl": "data/panel/model_crosswalk.csv",
      "encodingFormat": "text/csv",
      "sha256": "173b2e66fdc1ce1d2decd869553fb3f40e985deb8c2bb53693ed0b4430130c4f"
    },
    {
      "@type": "cr:FileObject",
      "@id": "codebook.md",
      "name": "codebook.md",
      "description": "Column-by-column codebook for the panel.",
      "contentUrl": "data/panel/codebook.md",
      "encodingFormat": "text/markdown",
      "sha256": "f025c3d8322afbcc6fe513e226a9a409f4fdb970a5e1b901ae8e04448423bc8d"
    }
  ],
  "recordSet": [
    {
      "@type": "cr:RecordSet",
      "@id": "panel",
      "name": "panel",
      "description": "Posted prices with provenance; gaps recorded as gaps, never interpolated.",
      "field": [
        {
          "@type": "cr:Field",
          "@id": "panel/model_id",
          "dataType": "sc:Text",
          "source": {
            "fileObject": {
              "@id": "panel.parquet"
            },
            "extract": {
              "column": "model_id"
            }
          }
        },
        {
          "@type": "cr:Field",
          "@id": "panel/month",
          "dataType": "sc:Text",
          "source": {
            "fileObject": {
              "@id": "panel.parquet"
            },
            "extract": {
              "column": "month"
            }
          }
        },
        {
          "@type": "cr:Field",
          "@id": "panel/price_blend_baseline",
          "dataType": "sc:Float",
          "source": {
            "fileObject": {
              "@id": "panel.parquet"
            },
            "extract": {
              "column": "price_blend_baseline"
            }
          }
        },
        {
          "@type": "cr:Field",
          "@id": "panel/price_flag",
          "dataType": "sc:Text",
          "source": {
            "fileObject": {
              "@id": "panel.parquet"
            },
            "extract": {
              "column": "price_flag"
            }
          }
        },
        {
          "@type": "cr:Field",
          "@id": "panel/source_url",
          "dataType": "sc:URL",
          "source": {
            "fileObject": {
              "@id": "panel.parquet"
            },
            "extract": {
              "column": "source_url"
            }
          }
        }
      ]
    }
  ]
}
Downloads last month
119

Paper for louisyzhu/price-of-intelligence