Orca Math Word Problems — LLM Reasoning Dataset
200K diverse grade-school math word problems synthetically generated by GPT-4 using an agent-based approach. Each problem is uniquely crafted without duplication from existing datasets, achieving state-of-the-art results with small 7B fine-tunes.
Dataset Details
| Provider | microsoft |
| Category | Reasoning |
| Size | 200K Problems |
| License | MIT |
| Downloads | 580k |
| Tags | Math, Synthetic, GPT-4, Word Problems, Microsoft |
from datasets import load_dataset
ds = load_dataset("microsoft/orca-math-word-problems-200k")
Fine-tune with this dataset
Estimated VRAM to fine-tune with QLoRA (4-bit base model + LoRA adapters), using conservative defaults:
| 7B QLoRA | ~6GB VRAM |
| 13B QLoRA | ~10GB VRAM |
Check if your GPU can fine-tune this →
New to fine-tuning? Follow the step-by-step walkthrough: Fine-Tune Your First LLM in 1 Hour
Related datasets
- MetaMathQA — Boosting GSM8K/MATH-style math skills in 7B models
- NuminaMath — Chain-of-thought math fine-tuning up to olympiad level
- OpenThoughts3-1.2M — Distilling strong math/code/science reasoning into 7B–32B models
- s1K-1.1 — Cheap, fast reasoning fine-tunes — 1k samples means minutes of training, not days
Frequently asked questions
Can I use Orca Math Word Problems commercially?
Yes — Orca Math Word Problems is released under MIT, a permissive license that allows commercial use, including training models you ship in a product. Check the dataset card for attribution requirements before release.
How much data does Orca Math Word Problems contain, and do I need all of it?
Orca Math Word Problems contains 200K Problems. You rarely need all of it: for style and format fine-tuning, a few hundred to a few thousand examples are enough — load a slice (e.g. split="train[:1000]") and scale up only if quality plateaus.
What is Orca Math Word Problems best used for?
Grade-school math word problems for small models. It belongs to the Reasoning section of our dataset hub, where you'll find alternatives and complementary sets.
← All datasets | Fine-Tuning Guide