Ploba logo

Find the right AI tool for the job: apps, agents, APIs, MCP servers and datasets, explained plainly.

Get the AI Digest

Discover

  • Explore
  • Top Charts
  • Categories
  • Agents
  • MCP servers
  • Developer tools

Learn & community

  • Learn AI
  • AI Glossary
  • Blog
  • Hackathons
  • People
  • AI Digest

For makers

  • List your tool
  • Docs
  • Claim a listing
  • Advertise
  • Write for us

Company

  • About
  • Careers
  • Changelog
  • Trust
  • Security
  • Contact

© 2026 Ploba. All rights reserved.

  • Privacy
  • Terms
  • Cookies
  • Advertising
  • All policies
llms.txt
Ploba logo
DiscoverHackathonsPeopleLearnAI Digest
Sign inGet started
HomeTop ChartsExploreCategories
TypesWeb apps1,894Agents585Plugins416APIs376Mobile apps152Datasets85
AdvertisingYour product hereReach people browsing AI tools on Ploba.Advertise hereAdvertisingYour product hereReach people browsing AI tools on Ploba.Advertise here
  1. Explore
  2. /
  3. Data
  4. /
  5. GSM8K
Data
GSM8K icon

GSM8K

by OpenAI

GSM8K is OpenAI's dataset of about 8.5K grade-school math word problems with step-by-step natural-language solutions. It is for researchers testing and training language models on multi-step arithmetic reasoning.

Dataset·Free

View on Hugging Face(opens in a new tab)

About GSM8K

Key features of GSM8K

  • About 8.5K grade-school math word problems
  • Step-by-step natural-language solutions with calculator annotations
  • Final numeric answer marked with ####
  • socratic configuration with guiding sub-questions
  • 7,473 training and 1,319 test problems
  • MIT License

Who GSM8K is for

  • LLM researchers benchmarking math reasoning
  • Teams fine-tuning models for chain-of-thought reasoning
  • Evaluation and leaderboard maintainers
  • Students studying language model reasoning

How to use GSM8K

  1. 1

    Step 1: Install the Hugging Face datasets library with pip install datasets.

  2. 2

    Step 2: Load the data with load_dataset("openai/gsm8k", "main"), or "socratic" for the sub-question version.

  3. 3

    Step 3: Use the train split for fine-tuning or few-shot examples and the test split for evaluation.

  4. 4

    Step 4: Extract the final answer after #### in each solution and compare it with your model's answer.

GSM8K FAQ

How many problems are in GSM8K?

About 8.5K: 7,473 in the training split and 1,319 in the test split.

What license is GSM8K released under?

The GSM8K dataset is licensed under the MIT License.

What is the difference between GSM8K main and socratic?

Both have the same problems and splits. The socratic configuration adds a guiding sub-question before each reasoning step in the answer.

What language is GSM8K in?

English.

Details

Pricing
FreeFree to use under the MIT License.
Type
Dataset
Categories
DataEducationResearch
Added
Sep 17, 2026
OpenAI iconPublisherOpenAI

Built from public sources

We put this page together from GSM8K’s own website and docs so you can compare it with other tools. Ploba isn’t affiliated with GSM8K.

Work at GSM8K? Claim the listing to manage it yourself.

Something wrong here?

More from OpenAI

View publisher
OpenAI Dots iconOpenAI DotsOpenAI · SubscriptionDots are always-on agents in ChatGPT that take on ongoing responsibilities and keep making progress between conversations. You choose which apps they use and when they need your approval.Productivity · SubscriptionChatGPT Work iconChatGPT WorkOpenAI · FreemiumChatGPT Work is an agent in ChatGPT that takes action across your apps and files and turns a goal into finished work such as sheets, slides, docs and web apps. It can stay with a project for hours.Productivity · FreemiumChatGPT Space iconChatGPT SpaceOpenAI · SubscriptionChatGPT Space is a shared place in ChatGPT for pages, files, spreadsheets and slides that you build with ChatGPT and your teammates. It replaces Library and is available on Pro, Business and Enterprise plans.Productivity · Subscription

Related tools

See more
WDWikipedia DatasetWikimedia · FreeThe Wikimedia Wikipedia dataset holds cleaned Wikipedia articles with one subset per language, built from the Wikipedia dumps.Data · FreeSQSQuADPranav Rajpurkar · FreeSQuAD (Stanford Question Answering Dataset) is a reading comprehension dataset of crowdworker questions on Wikipedia articles, where each answer is a span of text from the passage. It is for researchers training and evaluating extractive question-answering models.Data · FreeMMMMLUCAIS · FreeMMLU is a multiple-choice benchmark covering 57 subjects, from elementary mathematics to law, published by CAIS on Hugging Face. It measures how much knowledge a language model has across many fields.Data · Free