Off-the-shelf Datasets
Toloka's catalogue of ready-made training datasets sold outright — three named at the time of writing (Tau-bench Dataset Extension, University-level Math Reasoning, Multimodal Conversations) — as distinct from the custom data work its Platform sells.
Find alternatives to Off-the-shelf Datasets
What it does
- Data labelling
Used for
- Evaluation and observability
Features 9
- Enterprise RL environmentsRL environments where agents complete multi-step work across enterprise tools in 11 industry environments.
- Long-Horizon Knowledge WorkTasks producing slide decks, financial models and legal memos graded against expert answers.
- Terminal-Bench 2.1 ExtensionPrivate terminal and CLI agent tasks that are deterministic and validated against false results.
- Mobile Use RL GymsUI-fidelity replicas of real mobile apps as self-contained RL packages graded by state-diff.
- Household activity footageFirst-person household footage for robotics pre-training and imitation learning with time-aligned annotation.
- ToolathlonSimulated companies with search-only knowledge bases, discoverable tools and a user who withholds facts.
- Frontier STEMOffers PhD-authored problems across seven domains, weighted toward math and physics.
- SciCode ExtensionOffers expert-authored multi-step scientific coding tasks.
- EnterpriseBench KnowledgeSimulates whole companies with knowledge bases reachable only by search.
How it compares
-
Surge AI data platform
Toloka offers global multilingual crowdsourcing with pay-as-you-go pricing, which can undercut Surge on commodity tasks.
Research report · 19 Sep 2026
Sources
- Pricing
-
Official documentation · 5 Sep 2026
'These datasets are available for purchase. Talk to an expert' — the only commercial line on the page; no figure anywhere on the site.
- Description
-
Official documentation · 5 Sep 2026
Fetched 2026-09-05, HTTP 200, title 'Datasets ∙ Toloka', H1 'Datasets'. 'High-quality datasets are the foundation for training AI and LLMs… These datasets are available for purchase.' Three datasets listed by name. The footer navigation labels the section 'Off-the-shelf Datasets', which is the name used here; the page H1 is the bare 'Datasets'.
- Used for
-
Official documentation · 24 Oct 2026
'Evaluation, SFT and RL with verifiable rewards.' (https://toloka.ai/ots-datasets)
- Sold within
-
Live re-check 2026-09-15: toloka.ai/datasets/ states only that "these datasets are available for purchase" — does not clarify whether a plan or platform access is required either way.
- Intended for
-
datasets aimed at "developers working on robotics models, text and reasoning capabilities, and coding models" and "researchers focused on responsible AI and safety evaluation"
Also from Toloka
-
Toloka Platform Data service
A self-serve platform where an AI agent turns a described data goal into a full human-annotation pipeline - RLHF and preference data, data collection, instruction tuning, model evaluation, synthetic-data validation and content-moderation QA - with LLM-based quality checks on the output.
-
Toloka Arena Platform
An independent evaluation platform that ranks frontier LLMs on agentic tool-use tasks using private, non-contaminated benchmarks across industry domains, scored on a pass^5 reliability metric, with the underlying RL Gyms and evaluation datasets available to license.
-
Toloka Train Platform
A self-serve service that lowers per-request inference cost through two tools - fine-tuning LoRA adapters on frozen Qwen3 base models to replace a frontier API on a narrow task, and prompt gisting that compresses long instruction prefixes into learned tokens.
-
Tendem AI agent
A hybrid AI-plus-human agent that takes a delegated task - research, data analysis, copywriting, design or development - has AI do the first pass, then routes it to one of 10,000+ vetted experts for verification and multi-layer QA, returning results in 2-24 hours.
-
Toloka Physical AI Data service
Human-in-the-loop data programs -- demonstrations, annotation and evaluation -- for training robotics and physical AI systems.
Something wrong here? Send a correction — quote this product id: toloka-datasets.