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. AI
  4. /
  5. Nemotron Math Proofs v3 SFT
AI
Nemotron Math Proofs v3 SFT icon

Nemotron Math Proofs v3 SFT

by NVIDIA

Nemotron Math Proofs v3 SFT is an NVIDIA dataset of long-form proof generation, refinement, verification and meta-verification traces. It is for training language models on mathematical reasoning.

Dataset·Free

View on Hugging Face(opens in a new tab)

About Nemotron Math Proofs v3 SFT

Key features of Nemotron Math Proofs v3 SFT

  • 414,890 samples across 15,818 unique problems
  • Four task types: proof, refinement, verification and meta-verification
  • Multi-round refinement using prior attempts and verifier feedback
  • Verification scores of 0, 0.5 or 1
  • JSONL format, split into 10 train shards
  • Strict filtering of incomplete, malformed or length-capped generations
  • Task type identified by the metadata.data_type field

Who Nemotron Math Proofs v3 SFT is for

  • Teams training language models on proof writing
  • Researchers studying proof verification and self-verifying reasoning
  • Developers building theorem-proving systems with long-context reasoning

How to use Nemotron Math Proofs v3 SFT

  1. 1

    Step 1: Open the dataset page on Hugging Face and review the README.

  2. 2

    Step 2: Load the train split with the Hugging Face datasets library, which reads the 10 JSONL shards as one split.

  3. 3

    Step 3: Filter rows by metadata.data_type to pick proof, refinement, verification or meta_verification samples.

  4. 4

    Step 4: Use the messages field as supervised fine-tuning data for your model.

  5. 5

    Step 5: Credit the dataset according to the CC BY 4.0 license.

Nemotron Math Proofs v3 SFT FAQ

What license does the dataset use?

CC BY 4.0. NVIDIA says the dataset is ready for commercial or non-commercial use.

How do I tell the task types apart?

Use metadata.data_type, which is proof, refinement, verification or meta_verification. The subset field is only kept for compatibility.

How large is the dataset?

It has 414,890 samples from 15,818 unique problems, with a total disk size of about 122 GiB.

Where do the problems come from?

They are selected from the AoPS subset of an earlier Nemotron proofs dataset, using hard problems based on a prior pass-rate evaluation.

Details

Pricing
FreeFree under CC BY 4.0. NVIDIA states it can be used commercially or non-commercially.
Type
Dataset
Categories
AIDataEducation
Added
Sep 17, 2026
NVIDIA iconPublisherNVIDIA

Built from public sources

We put this page together from Nemotron Math Proofs v3 SFT’s own website and docs so you can compare it with other tools. Ploba isn’t affiliated with Nemotron Math Proofs v3 SFT.

Work at Nemotron Math Proofs v3 SFT? Claim the listing to manage it yourself.

Something wrong here?

More from NVIDIA

View publisher
Open-SWE Traces iconOpen-SWE TracesNVIDIA · FreeOpen-SWE-Traces is an NVIDIA dataset of agent trajectories for fixing real software issues, built for fine-tuning and distilling software engineering agents.Developer Tools · FreePhysicalAI Autonomous Vehicles iconPhysicalAI Autonomous VehiclesNVIDIA · FreePhysicalAI Autonomous Vehicles is a gated NVIDIA dataset for developing autonomous vehicle and automated driving systems, released under the NVIDIA Autonomous Vehicle Dataset License Agreement.AI · FreeNVIDIA Elements MCP iconNVIDIA Elements MCPNVIDIA · Contact for pricingNVIDIA Elements MCP connects AI coding assistants such as Claude Code, Cursor and Codex to NVIDIA's Elements design system, with component APIs, design tokens and examples. It is for developers building AI infrastructure, robotics and autonomous-vehicle interfaces.Developer Tools · Contact for pricing

Related tools

See more
FineWeb-Edu iconFineWeb-EduHugging Face · FreeFineWeb-Edu is a 1.3 trillion token dataset of educational web pages filtered from Hugging Face's FineWeb using an educational-quality classifier. It is for teams pretraining or researching large language models.AI · FreeMRMath ReasoningIFM · FreeMath-Reasoning from IFM is a dataset of math problem-solving, rewriting and dialogue data for training reasoning-oriented language models. It is one repository in the K2 Horizon dataset series.AI · FreeUMUltraData MathOpenBMB · FreeUltraData-Math is an OpenBMB math pre-training dataset of over 290B tokens in three quality tiers, built to improve mathematical reasoning in language models.AI · Free