Skip to datasets
Blobfish AI
BenchmarksTalk to usOpen app

Blobfish Research / Dataset factory

From public datasets
to executable work.

Explore the source catalog behind our benchmarks. Find MCP and domain datasets, inspect their classification evidence, and follow their path to a world with tools, tasks, assets, and trajectories.

Explore source datasets ↓View released benchmarks →
294Current Harbor listings
1,032,990Hugging Face sources inspected
146,818Licensed adaptation candidates

Coverage you can inspect.

Harbor: all 294 public listings observed; historical entries are also retained. Hugging Face: 1,032,988 rows checked through unfiltered pagination, alongside the earlier search inventory. The pagination pass is complete.

Related domain families provide working examples. A separate adaptation has not yet been built for each source row. Public access alone does not establish an open license.

Snapshot 2026-09-04 UTCDownload 5,000-row preview JSON ↗Full inspected inventory on Hugging Face ↗Browse the full Hugging Face catalog ↗

Available now

Inspect a complete benchmark.

HubBench contains 104 published tasks across 13 worlds, with shared state across CLI, website, API, and MCP tools. Its release receipts report 104/104 oracle passes. Current model results and historical pilots are labeled separately.

Leaderboard & qualification →Sample tasks →Assets & environment →Trajectories →Hugging Face ↗Harbor ↗World source on GitHub ↗

Source-specific release / Partial coverage

Arc CRM 6: inspect the adaptation.

Six independent synthetic workflows inspired by the pinned source; not 1,200 converted conversations or upstream API parity.

6/6 local and 6/6 registry Docker oracle passes. These establish solvability, not model performance.

All 6 tasks & assets →Actual reference trajectories →Pinned source assessment →Pinned Hugging Face release ↗Harbor package ↗World source on GitHub ↗

Source inventory

Find the next benchmark.

4,705 matching records

This website snapshot contains 5,000 of 1,033,285 inspected records, with Harbor listings and licensed candidates prioritized. Filters apply to the displayed snapshot.

Source datasets, relevance, license, and adaptation status
DatasetDomain & toolsLicenseAdaptation
Hugging FaceArsParadox/Viel-v1.0Upstream dataset ↗Manufacturing Engineering Design
Domain specific
mitWithin reuse policyReady for adaptationSource-specific world not built
Hugging FaceArsSocratica/egora-benchmarksUpstream dataset ↗Healthcare
Domain specific
agpl-3.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArseniy-Sandalov/Georgian-Sentiment-AnalysisUpstream dataset ↗Scientific Research
Domain specific
mitWithin reuse policyReady for adaptationSource-specific world not built
Hugging FaceArseniy-Sandalov/Med-NL2SQLUpstream dataset ↗Data Engineering Analytics
Domain specific
mitWithin reuse policyReady for adaptationSource-specific world not built
Hugging FaceArsh9210/Nemotron-3-Nano-RL-Training-BlendUpstream dataset ↗Policy Compliance Instruction Following
Domain specific
odc-byWithin reuse policyReady for adaptationSource-specific world not built
Hugging FaceArsh9210/Nemotron-Agentic-v1Upstream dataset ↗Customer Workplace Agents
MCP / tool useDomain specific
cc-by-4.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArsh9210/Nemotron-Competitive-Programming-v1Upstream dataset ↗Scientific Research
Domain specific
cc-by-4.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArsh9210/Nemotron-RL-Agentic-Conversational-Tool-Use-Pivot-v1Upstream dataset ↗Policy Compliance Instruction Following
MCP / tool useDomain specific
cc-by-4.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArsh9210/Nemotron-RL-Agentic-Function-Calling-Pivot-v1Upstream dataset ↗Policy Compliance Instruction Following
MCP / tool useDomain specific
cc-by-4.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArsh9210/Nemotron-RL-Instruction-Following-Calendar-v2Upstream dataset ↗Policy Compliance Instruction Following
Domain specific
cc-by-4.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArsh9210/Nemotron-RL-Instruction-Following-Citation-Formatting-v1Upstream dataset ↗Policy Compliance Instruction Following
Domain specific
cc-by-4.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArsh9210/Nemotron-RL-Instruction-Following-Free-Form-Formatting-v1Upstream dataset ↗Policy Compliance Instruction Following
Domain specific
cc-by-4.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArsh9210/Nemotron-RL-Instruction-Following-Structured-Outputs-v2Upstream dataset ↗Policy Compliance Instruction Following
Domain specific
cc-by-4.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArsh9210/Nemotron-RL-litmus-bench-v0.1Upstream dataset ↗Scientific Research
Domain specific
cc-by-4.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArsh9210/Nemotron-SFT-Agentic-v2Upstream dataset ↗Customer Workplace Agents
MCP / tool useDomain specific
apache-2.0, cc-by-4.0, mitWithin reuse policyReady for adaptationSource-specific world not built
Hugging FaceArsh9210/Nemotron-SFT-OpenCode-v1Upstream dataset ↗Unresolved
MCP / tool use
cc-by-4.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshR/test_dataset_1_20260821_141742Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshTandon/AAT_Deekfake_and_real_AudioUpstream dataset ↗Scientific Research
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/all_tasks_100_percent_mergedUpstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/all_tasks_10_percent_mergedUpstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/all_tasks_25_percent_mergedUpstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/all_tasks_50_percent_mergedUpstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/conditional-manipulation-AllUpstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/conditional-manipulation-task-1Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/conditional-manipulation-task-1_10_percentUpstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/conditional-manipulation-task-1_25_percentUpstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/conditional-manipulation-task-1_50_percentUpstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/conditional-manipulation-task-2Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/conditional-manipulation-task-2_10_percentUpstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/conditional-manipulation-task-2_25_percentUpstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/conditional-manipulation-task-2_50_percentUpstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/conditional-manipulation-task-3Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/conditional-manipulation-task-3_10_percentUpstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/conditional-manipulation-task-3_25_percentUpstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/conditional-manipulation-task-3_50_percentUpstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/conditional-manipulation-task-6Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/eval_demoUpstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/eval_pi05_test_1773695062Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/eval_pick-and-place-pi0-1773687641Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/eval_pick-and-place-pi05-1773693286Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/eval_pick-and-place-pi05-1773693739Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/eval_pick-and-place-pi05-1773694411Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/eval_pick-and-place-smolvla-1773670464Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/eval_pick-and-place-smolvla-1773673129Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/eval_pick-and-place-smolvla-1773674945Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/eval_pick-and-place-smolvla-1773676503Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/eval_pick-and-place-smolvla-1773865681Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/eval_pick-and-place-smolvla-1773865802Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/eval_pick-and-place-smolvla-1774573504Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built
Hugging FaceArshiaE/eval_pick-and-place-v10Upstream dataset ↗Manufacturing Engineering Design
Domain specific
apache-2.0Within reuse policyReady for adaptationSource-specific world not built