# OpenR1-Math-220k by Open R1: AI Training Dataset Source: https://savrn.com/datasets/openr1-math-220k Markdown alternate of the page above; the site index is https://savrn.com/llms.txt --- ## Dataset Card By Open R1, published under apache-2.0, revision e4e141ec9dea. ### Dataset description OpenR1-Math-220k is a large-scale dataset for mathematical reasoning. It consists of 220k math problems with two to four reasoning traces generated by [DeepSeek R1](https://savrn.com/models/deepseek-r1) for problems from NuminaMath 1.5. The traces were verified using [Math Verify](https://github.com/huggingface/Math-Verify) for most samples and [Llama-3.3-70B-Instruct](https://savrn.com/models/llama-3-3-70b-instruct) as a judge for 12% of the samples, and each problem contains at least one reasoning trace with a correct answer. The dataset consists of two splits: - default with 94k problems and that achieves the best performance after SFT. - extended with 131k samples where we add data sources like cn_k12. This provides more reasoning traces, but we found that the performance after SFT to be lower than the default subset, likely because the questions from cn_k12 are less difficult than other sources. You can load the dataset as follows: ``` from datasets import load_dataset ds = load_dataset("open-r1/OpenR1-Math-220k", "default") ``` ### Dataset curation [Read the full dataset card (306 words)](https://savrn.com/datasets/openr1-math-220k/card) ## Structure ### default 93,733 rows | Split | Rows | Size | | --- | --- | --- | | train | 93,733 | 5.1 GB | problemstringsolutionstringanswerstringproblem_typestringquestion_typestringsourcestringuuidstringis_reasoning_completeSequencegenerationsSequencecorrectness_math_verifySequencecorrectness_llamaSequencefinish_reasonsSequencecorrectness_countint64messageslist ### all 225,129 rows | Split | Rows | Size | | --- | --- | --- | | train | 225,129 | 10.2 GB | problemstringsolutionstringanswerstringproblem_typestringquestion_typestringsourcestringuuidstringis_reasoning_completeSequencegenerationsSequencecorrectness_math_verifySequencecorrectness_llamaSequencefinish_reasonsSequencecorrectness_countint64messageslist ### extended 131,396 rows | Split | Rows | Size | | --- | --- | --- | | train | 131,396 | 4.8 GB | problemstringsolutionstringanswerstringproblem_typestringquestion_typestringsourcestringuuidstringis_reasoning_completeSequencegenerationsSequencecorrectness_math_verifySequencecorrectness_llamaSequencefinish_reasonsSequencecorrectness_countint64messageslist ## Details Repository open-r1/OpenR1-Math-220k Publisher Open R1 Task category Not stated by the source Tags Not stated by the source Size category Not stated by the source Languages en Revision e4e141ec9dea9f8326f4d347be56105859b2bd68 Last updated 2025-02-18 ## Files 63 files, 12.6 GB in total. Data60 files · 12.6 GB Documentation1 file · 5.1 KB Repository2 files · 4.9 KB Every file | File | Type | Size | SHA-256 | | --- | --- | --- | --- | | all/default-00000-of-00010.parquet | Data | 213.6 MB | ccc3a95e4875 | | all/default-00001-of-00010.parquet | Data | 214.6 MB | a7d3eb54d97f | | all/default-00002-of-00010.parquet | Data | 214.6 MB | a44a2354bfa7 | | all/default-00003-of-00010.parquet | Data | 216.8 MB | dd7d2f0f04d8 | | all/default-00004-of-00010.parquet | Data | 214.8 MB | e358157990d7 | | all/default-00005-of-00010.parquet | Data | 214.5 MB | 30254d14d17b | | all/default-00006-of-00010.parquet | Data | 215.6 MB | b3bce93c06d4 | | all/default-00007-of-00010.parquet | Data | 216.3 MB | 7def8029cf6d | | all/default-00008-of-00010.parquet | Data | 213.8 MB | 29cf75a663e9 | | all/default-00009-of-00010.parquet | Data | 215.3 MB | 14df2775196a | | all/extended-00000-of-00010.parquet | Data | 207.5 MB | f9716086950b | | all/extended-00001-of-00010.parquet | Data | 204.5 MB | d96a9f53e68e | | all/extended-00002-of-00010.parquet | Data | 208.6 MB | c58b7dd82c4e | | all/extended-00003-of-00010.parquet | Data | 206.2 MB | 2acd199fa543 | | all/extended-00004-of-00010.parquet | Data | 205.3 MB | ebcef1d03a82 | | all/extended-00005-of-00010.parquet | Data | 205.2 MB | 6facabc93453 | | all/extended-00006-of-00010.parquet | Data | 207.7 MB | 9d31ebbaf724 | | all/extended-00007-of-00010.parquet | Data | 205.9 MB | 7d8f224f05f1 | | all/extended-00008-of-00010.parquet | Data | 206.8 MB | 1e67afba40f4 | | all/extended-00009-of-00010.parquet | Data | 206.2 MB | ceebbc82c991 | | all/train-00000-of-00020.parquet | Data | 210.9 MB | fc09ca74674f | | all/train-00001-of-00020.parquet | Data | 212.7 MB | 9c3aa4f99ccc | | all/train-00002-of-00020.parquet | Data | 211.2 MB | c6e1635c9739 | | all/train-00003-of-00020.parquet | Data | 209.7 MB | 9199725eba5e | | all/train-00004-of-00020.parquet | Data | 210.2 MB | 360bdb99acef | | all/train-00005-of-00020.parquet | Data | 212.5 MB | 8d0ddce8c036 | | all/train-00006-of-00020.parquet | Data | 211.1 MB | 4fc1aed02932 | | all/train-00007-of-00020.parquet | Data | 210.9 MB | 279f6ca64239 | | all/train-00008-of-00020.parquet | Data | 212.1 MB | eef061ae2462 | | all/train-00009-of-00020.parquet | Data | 210.9 MB | c99427cb7a4e | | all/train-00010-of-00020.parquet | Data | 211.9 MB | 5ead84ce4c97 | | all/train-00011-of-00020.parquet | Data | 212.6 MB | 9407f4646cbb | | all/train-00012-of-00020.parquet | Data | 210.2 MB | 1b8f1f54bf06 | | all/train-00013-of-00020.parquet | Data | 210.1 MB | b3dbf4d982aa | | all/train-00014-of-00020.parquet | Data | 209.1 MB | efcc1588926f | | all/train-00015-of-00020.parquet | Data | 210.8 MB | d532305ac755 | | all/train-00016-of-00020.parquet | Data | 212.1 MB | 0d5776a238ab | | all/train-00017-of-00020.parquet | Data | 212.5 MB | 1148d77261b9 | | all/train-00018-of-00020.parquet | Data | 211.1 MB | 44c0d751c344 | | all/train-00019-of-00020.parquet | Data | 209.3 MB | 578fa30d9685 | | data/train-00000-of-00010.parquet | Data | 213.6 MB | ccc3a95e4875 | | data/train-00001-of-00010.parquet | Data | 214.6 MB | a7d3eb54d97f | | data/train-00002-of-00010.parquet | Data | 214.6 MB | a44a2354bfa7 | | data/train-00003-of-00010.parquet | Data | 216.8 MB | dd7d2f0f04d8 | | data/train-00004-of-00010.parquet | Data | 214.8 MB | e358157990d7 | | data/train-00005-of-00010.parquet | Data | 214.5 MB | 30254d14d17b | | data/train-00006-of-00010.parquet | Data | 215.6 MB | b3bce93c06d4 | | data/train-00007-of-00010.parquet | Data | 216.3 MB | 7def8029cf6d | | data/train-00008-of-00010.parquet | Data | 213.8 MB | 29cf75a663e9 | | data/train-00009-of-00010.parquet | Data | 215.3 MB | 14df2775196a | | extended/train-00000-of-00010.parquet | Data | 207.5 MB | f9716086950b | | extended/train-00001-of-00010.parquet | Data | 204.5 MB | d96a9f53e68e | | extended/train-00002-of-00010.parquet | Data | 208.6 MB | c58b7dd82c4e | | extended/train-00003-of-00010.parquet | Data | 206.2 MB | 2acd199fa543 | | extended/train-00004-of-00010.parquet | Data | 205.3 MB | ebcef1d03a82 | | extended/train-00005-of-00010.parquet | Data | 205.2 MB | 6facabc93453 | | extended/train-00006-of-00010.parquet | Data | 207.7 MB | 9d31ebbaf724 | | extended/train-00007-of-00010.parquet | Data | 205.9 MB | 7d8f224f05f1 | | extended/train-00008-of-00010.parquet | Data | 206.8 MB | 1e67afba40f4 | | extended/train-00009-of-00010.parquet | Data | 206.2 MB | ceebbc82c991 | | README.md | Documentation | 5.1 KB | — | | .gitattributes | Repository | 2.5 KB | — | | extended/.gitattributes | Repository | 2.5 KB | — | ## License and Download License apache-2.0 Access No access gate [Download from Open R1](https://huggingface.co/datasets/open-r1/OpenR1-Math-220k) Released by Open R1 through its official repository on Hugging Face. [Read the license](https://www.apache.org/licenses/LICENSE-2.0). ## Models Trained on This Dataset - Trained on (disclosed)[reasoning-rlvr-sft](https://savrn.com/models/reasoning-rlvr-sft) ## Open R1 [All models and datasets](https://savrn.com/model-publishers/open-r1) ## Explore More - [The dataset directory](https://savrn.com/datasets) - [The model directory](https://savrn.com/models) - [Research papers](https://savrn.com/papers) ## Source - Repository metadata, read 2026-10-19. - [Hugging Face record](https://huggingface.co/datasets/open-r1/OpenR1-Math-220k) - [How the hub is built](https://savrn.com/model-hub/methodology)