Datasets:
Tasks:
Question Answering
Formats:
json
Languages:
English
Size:
< 1K
Tags:
web-search
company-search
agentic-search
information-retrieval
entity-retrieval
multi-constraint
case_key string | question string | family string | constraint_count int64 | constraints list | predicates dict | census_metadata dict | source_metadata dict | sort_order int64 | gold list |
|---|---|---|---|---|---|---|---|---|---|
investor__front_range__inv_insight_partners__e2013_15 | Which companies count Insight Partners among their investors, were founded between 2013 and 2015, and are headquartered along the Colorado Front Range? | investor | 3 | [
"count Insight Partners among their investors",
"were founded between 2013 and 2015",
"are headquartered along the Colorado Front Range"
] | {
"axes": [
"investor_backing",
"founding_era",
"geo_region"
],
"labels": {
"investor_backing": "inv_insight_partners",
"founding_era": "e2013_15",
"geo_region": "front_range"
}
} | {
"atom_sizes": [
123,
836,
2294
],
"drop_one_sizes": [
133,
21,
10
],
"answer_count": 4,
"answers_in_range": true,
"drop_one_ambiguous": true,
"atoms_large": true,
"ok": true
} | {
"combination": "geo_region+investor_backing+founding_era",
"coverage": {
"known_product": 0.9158
}
} | 1 | [
{
"entity_key": "name:75ffdc3fb5d1c1f99f7050cd",
"name": "Automox",
"domain": "automox.com",
"aliases": [
"Automox Inc."
],
"label_notes": null
},
{
"entity_key": "name:b96f93f7591a8aa4429ffd71",
"name": "Quantive",
"domain": "quantive.com",
"aliases": [
"Form... |
investor__pacific_nw__inv_bessemer_venture_partners__e2016_18 | Which companies are headquartered in the Pacific Northwest, count Bessemer Venture Partners among their investors, and were founded between 2016 and 2018? | investor | 3 | [
"are headquartered in the Pacific Northwest",
"count Bessemer Venture Partners among their investors",
"were founded between 2016 and 2018"
] | {
"axes": [
"geo_region",
"investor_backing",
"founding_era"
],
"labels": {
"geo_region": "pacific_nw",
"investor_backing": "inv_bessemer_venture_partners",
"founding_era": "e2016_18"
}
} | {
"atom_sizes": [
312,
804,
2673
],
"drop_one_sizes": [
119,
58,
15
],
"answer_count": 4,
"answers_in_range": true,
"drop_one_ambiguous": true,
"atoms_large": true,
"ok": true
} | {
"combination": "geo_region+investor_backing+founding_era",
"coverage": {
"known_product": 0.9158
}
} | 2 | [
{
"entity_key": "name:c806dea4637890b01cd2a0f1",
"name": "Ganaz, Inc.",
"domain": "ganaz.com",
"aliases": [
"Ganaz",
"Ganaz Inc"
],
"label_notes": null
},
{
"entity_key": "name:b98d253674beeae2c8eb87f1",
"name": "Kaskada",
"domain": "kaskada.com",
"aliases": [... |
investor__sea_hub__b__inv_sequoia_capital__multi_office | Which companies raised a Series B, count Sequoia Capital among their investors, operate offices in more than one city, and are headquartered in Southeast Asia? | investor | 4 | [
"raised a Series B",
"count Sequoia Capital among their investors",
"operate offices in more than one city",
"are headquartered in Southeast Asia"
] | {
"axes": [
"round_stage",
"investor_backing",
"expansion",
"geo_region"
],
"labels": {
"round_stage": "b",
"investor_backing": "inv_sequoia_capital",
"expansion": "multi_office",
"geo_region": "sea_hub"
}
} | {
"atom_sizes": [
84,
6385,
1230,
4006
],
"drop_one_sizes": [
215,
5,
10,
5
],
"answer_count": 3,
"answers_in_range": true,
"drop_one_ambiguous": true,
"atoms_large": true,
"ok": true
} | {
"combination": "geo_region+round_stage+investor_backing+expansion",
"coverage": {
"known_product": 0.911
}
} | 3 | [
{
"entity_key": "name:6ce88345db2b7b3f6b68d8ba",
"name": "90 Seconds",
"domain": "90seconds.com",
"aliases": [
"90 Seconds Pte Ltd"
],
"label_notes": null
},
{
"entity_key": "name:2db7473baf2bf105c95c28ea",
"name": "99.co",
"domain": "99.co",
"aliases": [
"99 ... |
accelerator__sea_hub__seed__ai__yc | Which companies went through Y Combinator, are headquartered in Southeast Asia, raised a seed round, and build artificial-intelligence products? | accelerator | 4 | [
"went through Y Combinator",
"are headquartered in Southeast Asia",
"raised a seed round",
"build artificial-intelligence products"
] | {
"axes": [
"accelerator",
"geo_region",
"round_stage",
"industry"
],
"labels": {
"accelerator": "yc",
"geo_region": "sea_hub",
"round_stage": "seed",
"industry": "ai"
}
} | {
"atom_sizes": [
57,
5418,
3009,
5543
],
"drop_one_sizes": [
943,
8,
26,
5
],
"answer_count": 3,
"answers_in_range": true,
"drop_one_ambiguous": true,
"atoms_large": true,
"ok": true
} | {
"combination": "geo_region+round_stage+industry+accelerator",
"coverage": {
"known_product": 0.7468
}
} | 4 | [
{
"entity_key": "name:1aaed054b872b8fa13088838",
"name": "Axross",
"domain": "axross.co",
"aliases": [],
"label_notes": null
},
{
"entity_key": "name:94e035dfa7d946144ca46154",
"name": "Bot MD",
"domain": "botmd.com",
"aliases": [
"5 Health",
"5 Health Inc.",
... |
accelerator__front_range__pre_seed__science_eng__yc | Which companies went through Y Combinator, are headquartered along the Colorado Front Range, raised a pre-seed round, and work in science or engineering? | accelerator | 4 | [
"went through Y Combinator",
"are headquartered along the Colorado Front Range",
"raised a pre-seed round",
"work in science or engineering"
] | {
"axes": [
"accelerator",
"geo_region",
"round_stage",
"industry"
],
"labels": {
"accelerator": "yc",
"geo_region": "front_range",
"round_stage": "pre_seed",
"industry": "science_eng"
}
} | {
"atom_sizes": [
150,
6074,
3393,
5543
],
"drop_one_sizes": [
1601,
5,
9,
22
],
"answer_count": 3,
"answers_in_range": true,
"drop_one_ambiguous": true,
"atoms_large": true,
"ok": true
} | {
"combination": "geo_region+round_stage+industry+accelerator",
"coverage": {
"known_product": 0.7468
}
} | 5 | [
{
"entity_key": "name:7b84cc4655a06af24c9bd8c3",
"name": "Albedo",
"domain": "albedo.com",
"aliases": [
"Albedo Space",
"Albedo Space Corp.",
"AlbedoSpaceCorp"
],
"label_notes": null
},
{
"entity_key": "name:cc4469940384e956cc7b18d7",
"name": "Alt-X",
"domai... |
accelerator__canada_east__a__ai__yc | Which companies are headquartered in eastern Canada, raised a Series A, build artificial-intelligence products, and went through Y Combinator? | accelerator | 4 | [
"are headquartered in eastern Canada",
"raised a Series A",
"build artificial-intelligence products",
"went through Y Combinator"
] | {
"axes": [
"geo_region",
"round_stage",
"industry",
"accelerator"
],
"labels": {
"geo_region": "canada_east",
"round_stage": "a",
"industry": "ai",
"accelerator": "yc"
}
} | {
"atom_sizes": [
213,
1609,
3009,
5543
],
"drop_one_sizes": [
310,
19,
18,
6
],
"answer_count": 3,
"answers_in_range": true,
"drop_one_ambiguous": true,
"atoms_large": true,
"ok": true
} | {
"combination": "geo_region+round_stage+industry+accelerator",
"coverage": {
"known_product": 0.7468
}
} | 6 | [
{
"entity_key": "name:42b2f90d6656e319d3a01bea",
"name": "AON3D",
"domain": "aon3d.com",
"aliases": [
"MADE IN 3D CANADA INC.",
"aon3d Inc"
],
"label_notes": null
},
{
"entity_key": "name:00f6af807951b33000117c4f",
"name": "Aviron",
"domain": "avironactive.com",
... |
accelerator__canada_east__eight_figure__e2013_15__yc | Which companies are headquartered in eastern Canada, raised a single round of at least $10 million, were founded between 2013 and 2015, and went through Y Combinator? | accelerator | 4 | [
"are headquartered in eastern Canada",
"raised a single round of at least $10 million",
"were founded between 2013 and 2015",
"went through Y Combinator"
] | {
"axes": [
"geo_region",
"funding_class",
"founding_era",
"accelerator"
],
"labels": {
"geo_region": "canada_east",
"funding_class": "eight_figure",
"founding_era": "e2013_15",
"accelerator": "yc"
}
} | {
"atom_sizes": [
213,
1371,
1129,
5543
],
"drop_one_sizes": [
162,
16,
19,
5
],
"answer_count": 3,
"answers_in_range": true,
"drop_one_ambiguous": true,
"atoms_large": true,
"ok": true
} | {
"combination": "geo_region+funding_class+founding_era+accelerator",
"coverage": {
"known_product": 0.7865
}
} | 7 | [
{
"entity_key": "name:42b2f90d6656e319d3a01bea",
"name": "AON3D",
"domain": "aon3d.com",
"aliases": [
"MADE IN 3D CANADA INC.",
"aon3d Inc"
],
"label_notes": null
},
{
"entity_key": "name:62ecf26fe46ab00a547cd51a",
"name": "BenchSci",
"domain": "benchsci.com",
... |
accelerator__texas__priced_institutional__e2016_18__techstars | Which companies were founded between 2016 and 2018, went through Techstars, are headquartered in one of the large Texas metros, and raised at least one priced institutional round? | accelerator | 4 | [
"were founded between 2016 and 2018",
"went through Techstars",
"are headquartered in one of the large Texas metros",
"raised at least one priced institutional round"
] | {
"axes": [
"founding_era",
"accelerator",
"geo_region",
"funding_class"
],
"labels": {
"founding_era": "e2016_18",
"accelerator": "techstars",
"geo_region": "texas",
"funding_class": "priced_institutional"
}
} | {
"atom_sizes": [
255,
1643,
1876,
4711
],
"drop_one_sizes": [
198,
40,
19,
10
],
"answer_count": 3,
"answers_in_range": true,
"drop_one_ambiguous": true,
"atoms_large": true,
"ok": true
} | {
"combination": "geo_region+funding_class+founding_era+accelerator",
"coverage": {
"known_product": 0.7865
}
} | 8 | [
{
"entity_key": "name:6e5e96e23bb94eac617d579d",
"name": "Crave Retail",
"domain": "craveretail.com",
"aliases": [],
"label_notes": null
},
{
"entity_key": "name:07d7716749e988d8f5551296",
"name": "DocStation",
"domain": "docstation.co",
"aliases": [],
"label_notes": null... |
accelerator__latam_hub__seed_stage__e2013_15__yc | Which companies went through Y Combinator, are headquartered in a large Latin American hub, raised money at the seed stage, and were founded between 2013 and 2015? | accelerator | 4 | [
"went through Y Combinator",
"are headquartered in a large Latin American hub",
"raised money at the seed stage",
"were founded between 2013 and 2015"
] | {
"axes": [
"accelerator",
"geo_region",
"funding_class",
"founding_era"
],
"labels": {
"accelerator": "yc",
"geo_region": "latam_hub",
"funding_class": "seed_stage",
"founding_era": "e2013_15"
}
} | {
"atom_sizes": [
135,
8772,
1129,
5543
],
"drop_one_sizes": [
512,
5,
90,
5
],
"answer_count": 4,
"answers_in_range": true,
"drop_one_ambiguous": true,
"atoms_large": true,
"ok": true
} | {
"combination": "geo_region+funding_class+founding_era+accelerator",
"coverage": {
"known_product": 0.7865
}
} | 9 | [
{
"entity_key": "name:0858dcc7ede885a8e034793a",
"name": "99minutos",
"domain": "99minutos.com",
"aliases": [],
"label_notes": null
},
{
"entity_key": "name:57e03ac18a5704c27e9d300c",
"name": "AgendaPro",
"domain": "agendapro.com",
"aliases": [],
"label_notes": null
},
... |
funding__uk_ex_london__nine_figure__ai | Which companies build artificial-intelligence products, are headquartered in the UK outside London, and raised a single round of at least $100 million? | funding | 3 | [
"build artificial-intelligence products",
"are headquartered in the UK outside London",
"raised a single round of at least $100 million"
] | {
"axes": [
"industry",
"geo_region",
"funding_class"
],
"labels": {
"industry": "ai",
"geo_region": "uk_ex_london",
"funding_class": "nine_figure"
}
} | {
"atom_sizes": [
1826,
3047,
6066
],
"drop_one_sizes": [
197,
162,
65
],
"answer_count": 4,
"answers_in_range": true,
"drop_one_ambiguous": true,
"atoms_large": true,
"ok": true
} | {
"combination": "geo_region+funding_class+industry",
"coverage": {
"known_product": 0.981
}
} | 10 | [
{
"entity_key": "name:0c81fe44cc6a6ff9d44f15c2",
"name": "Connex One",
"domain": "connexone.io",
"aliases": [],
"label_notes": null
},
{
"entity_key": "name:851049654be3dcffaeaf7c58",
"name": "ConnexAI",
"domain": "connex.ai",
"aliases": [],
"label_notes": null
},
{
... |
OB Company Websearch
10 multi-constraint company-discovery questions with a frozen, hand-labelled reference set. The public set accompanies the OpenBenchmarks Multi Turn Company Search Benchmark.
The benchmark holds the research agent and budgets fixed while varying the web search provider. This release is for the search-only condition: the agent can search and use result snippets but is not given a page-fetch tool.
Dataset contents
Each question asks for the complete set of companies satisfying three or four simultaneous constraints. This release contains:
- 10 questions
- 3 three-constraint questions and 7 four-constraint questions
- 103 reference-set memberships across the questions
- 100 unique canonical companies
Each row contains:
case_key— stable question identifier.question— natural-language company-discovery request shown to the agent.family— broad construction family such as investor, accelerator, or funding.constraint_countandconstraints— number and text of the simultaneous conditions.predicates— normalized axes and labels used to construct the question.census_metadata— dataset-construction checks and reference-set size.source_metadata— construction coverage metadata.sort_order— stable display order.gold— reference companies, including canonical name, domain, aliases, and notes.
Dataset-level release metadata and the source content hash are retained in
metadata.json.
Evaluation
A fixed agent is asked each question and must return the complete matching
company set. Returned companies are resolved to canonical identities and
compared deterministically with gold:
- a reference company returned by the agent is a true positive;
- an extra company is a false positive;
- a missed reference company is a false negative.
Precision, recall, F1, and exact-set accuracy are computed per agent run. Each provider configuration is run three times on the same locked questions; reported quality metrics are the mean and sample standard deviation across those three runs.
This repository contains questions and labels, not vendor responses or benchmark scores.
Gold construction
Questions were selected from broad intersections in the source company snapshot. Candidate labels were reviewed in a DB-free local labeller and finalized only after every case was marked reviewed and the local gold release was published.
Reproducibility
The open runner and offline judge are available at openbenchmarks-labs/multi-turn-company-search. The benchmark methodology and current results are published at openbenchmarks.com/multi-turn-company-search.
Release: ob-company-websearch-v1
Schema: company-research-dataset-v1
Gold version: 1
Content SHA-256: sha256:b987888daa5e0516ac7cf18deaba3830496929fa653d2cb68b494bd2945664ab
- Downloads last month
- 15