{ "fetched_at": "2026-10-24T18:07:06Z", "source_url": "https://www.kaggle.com/competitions/ai-mathematical-olympiad-progress-prize-3/data", "pages": [ { "name": "rules", "content": "##ENTRY IN THIS COMPETITION CONSTITUTES YOUR ACCEPTANCE OF THESE OFFICIAL COMPETITION RULES.\n\n**[See Section 3.18 for defined terms](rules#18.-terms)**\n\n*The Competition named below is a skills-based competition to promote and further the field of data science. You must register via the Competition Website to enter. To enter the Competition, you must agree to these Official Competition Rules, which incorporate by reference the provisions and content of the Competition Website and any Specific Competition Rules herein (collectively, the \"Rules\"). Please read these Rules carefully before entry to ensure you understand and agree. You further agree that Submission in the Competition constitutes agreement to these Rules. You may not submit to the Competition and are not eligible to receive the prizes associated with this Competition unless you agree to these Rules. These Rules form a binding legal agreement between you and the Competition Sponsor with respect to the Competition. Your competition Submissions must conform to the requirements stated on the Competition Website. Your Submissions will be scored based on the evaluation metric described on the Competition Website. Subject to compliance with the Competition Rules, Prizes, if any, will be awarded to Participants with the best scores, based on the merits of the data science models submitted. See below for the complete Competition Rules. For Competitions designated as hackathons by the Competition Sponsor (“Hackathons”), your Submissions will be judged by the Competition Sponsor based on the evaluation rubric set forth on the Competition Website (“Evaluation Rubric”). The Prizes, if any, will be awarded to Participants with the highest ranking(s) as determined by the Competition Sponsor based on such rubric.*\n\n**You cannot sign up to Kaggle from multiple accounts and therefore you cannot enter or submit from multiple accounts.**\n\n

1. COMPETITION-SPECIFIC TERMS

\n

1. COMPETITION TITLE

AI Mathematical Olympiad - Progress Prize 3\n

2. COMPETITION SPONSOR

XTX Investments Limited \n

3. COMPETITION SPONSOR ADDRESS

R7, 14-18 Handyside Street, London N1C 4DN, UK\n

4. COMPETITION WEBSITE

https://www.kaggle.com/competitions/ai-mathematical-olympiad-progress-prize-3\n

5. TOTAL PRIZES AVAILABLE: $2,207,152

\n\n**Overall Progress Prize Winner**: The Overall Progress Prize Winner shall be the highest ranking team that first achieves a score of at least 47/50 on both public and private test sets. After any prizes for the five top-ranking teams have been awarded, the remainder of the total fund shall be awarded to the Overall Progress Prize Winner.\n\nIf a team is named the Overall Progress Prize Winner in this competition, the prize will be at least $1,589,248. If no team is named the Overall Progress Prize Winner in this competition, the remainder of the total fund shall roll over to the next competition, where the same prize allocation will apply.\n\n##**TOTAL FUND FOR PROGRESS PRIZE 3: $2,207,152**\n\n###**Prizes for Top-Ranking Teams in this Competition**:\n- 1st Place: $262,144\n- 2nd Place: $131,072\n- 3rd Place: $65,536\n- 4th Place: $32,768\n- 5th Place: $16,384\n\n**Overall Progress Prize Winner**: The Overall Progress Prize Winner shall be the highest ranking team that achieves a score of at least 47/50 on both public and private test sets. After any prizes for the five top-ranking teams have been awarded, the remainder of the total fund shall be awarded to the Overall Progress Prize Winner.\n\nIf a team is named the Overall Progress Prize Winner in this competition, the prize will be at least $1,589,248. If no team is named the Overall Progress Prize Winner in this competition, the remainder of the total fund shall roll over to the next competition, where the same prize allocation will apply.\n\n**Additional Prizes: $110,000**\n\n- **Longest Leader Prize**: $20,000\n- **Hard Problem Prize**: $30,000\n- **Math Corpus Prize**: $30,000\n- **Writeup Prizes**: 2x $15,000 \n\nLongest Leader Prize: $20,000\nAwarded to the team who's public notebook(s) generates the best scoring submission on the leaderboard for the longest period of time between November 20, 2025 and February 2, 2026 11:59 PM UTC. The public notebooks need to adhere to the same requirements and restrictions regarding licensing, reproducibility, and documentation to which the winning Submission is subject (see [Competition Rules](https://www.kaggle.com/competitions/ai-mathematical-olympiad-progress-prize-3/rules)). The notebook(s) and any datasets used must be made publicly available at latest by February 9, 2026 11:59 PM UTC and kept public until the final Progress Prizes are awarded to the winning Teams at the end of the competition. In the event that a team has multiple public notebooks, each achieving the top leaderboard position at different times, those durations will be combined.\n\nHard Problem Prize: $30,000\nAwarded to the highest ranked team(s) on the private leaderboard who solved the most difficult problem(s) in the private test set. Problem difficulty is measured by the average accuracy score across all selected submissions at the end of the competition, where average accuracy is calculated as described in the Evaluation section (accounting for both runs of each submission). For example, if the most difficult problem has an average accuracy of 1.7%, then the highest ranked team on the private leaderboard who achieved a full score on that problem will receive the prize. If multiple problems tie for \"most difficult\": We will identify the highest ranked team for each tied problem separately, and the $30,000 prize will be split equally among the winning teams. For example, if two problems tie for most difficult, each with its own highest-scoring solver, the prize will be divided into two $15,000 awards. In the event of a tie among notebooks, the tiebreaker will be based on submission time (first to submit).\n\nMath Corpus Prize: $30,000 \nAwarded to the team who provides the most valuable dataset to the community for improving AIMO model performance.\n\n*Math Corpus Prize Requirements*\n- The dataset must be publicly released either on Kaggle or on HuggingFace prior to February 9, 2026 11:59 PM UTC. Teams must create a Kaggle Discussion post until this date that explicitly links to their dataset and tags it as an entry for the Math Corpus Prize. All eligible datasets will be judged according to the criteria below, and the highest scoring dataset will be awarded the prize.\n - Your Kaggle Discussion post will be used to judge your dataset submission based on the criteria below, and your claims will be independently verified by the Host Team.\n- The dataset must be in English.\n- The latest timestamped version of the dataset will be used for evaluation.\n- The dataset is not allowed to exceed 5M datapoints, which is an upper bound on common datasets used by prior AIMO winners to fine-tune their models. Each datapoint is not allowed to exceed more than 100k characters. \n- Each datapoint from the submitted dataset must come with an open source license that allows free dissemination of data.\n\n*Math Corpus Prize Evaluation Criteria*:\n\n- Data Novelty (25 points) \n - Is the dataset distinctly unique from other datasets found on the internet? Minor modification are not considered novel. In particular, translations into English from other languages are not considered significant modifications.\n - Describe what methods you used to ensure your dataset is novel, and note how your dataset does not overlap with prior submitted datasets.\n- Format (25 points) \n - Does the dataset comes in a format that makes it easy to handle and train a model?\n - Does each datapoint contain rich metadata?\n- Performance (50 points) \n - Does the dataset improve mathematical reasoning and aid in improving model performance during the AI Mathematical Olympiad. How so and how much?\n\n#### Writeup Prize: $30,000 (2 awards x $15,000)\nAwarded to the two best Solution Writeups from non-winning teams based on the below criteria.\n\n*Writeup Prize Requirements*\n- The Writeup should be an official [Kaggle Writeup](https://www.kaggle.com/discussions/product-announcements/593763), attached to your submission within one week after the competition ended.\n- It should mirror an academic publication about your submission, and include graphs that outline how your model was trained, the impact of different techniques and methods on performance, and how it could be further improved. (The arXiv post by the AIMO2 winners is a good reference paper in terms of academic standard: https://arxiv.org/abs/2504.16891)\n- It should be limited to 30k words (which approximately corresponds to a 50-page paper), and split into a main section of not more than 5k words (approximately 10 pages), which outlines your methodology, and the remaining sections, which include the bibliography, examples, and any other materials that support the main section.\n\n*Writeup Prize Evaluation Criteria*\n- Clarity (10 points)\n - Does the Writeup highlight in detail the entire model creation lifecycle (including data collection, training, and benchmarking), and potentially variants of the model that you created?\n- Ablation Studies and Variants (10 points)\n - Does the Writeup contain information that pinpoints where the performance gains of your model came from? \n- Comparison with the State of the Art (10 points)\n - Does the Writeup evaluate how your model (potentially split across types of reasoning modes or mathematical problems) compares to the state of the art for both open-weight and commercial models?\n- Graphs and Charts (10 points) \n - Does the Writeup contain well-designed visuals that clearly \"tell the story\" by illustrating your model's performance, training process, and key experimental results?- Reproducibility (60 points) \n - Is your write-up sufficiently detailed to easily allow the reader to reproduce the training of your model starting from the same artifacts (base open-weight models and data) that you started from? \n\n\n

6. WINNER LICENSE TYPE

CC-BY 4.0\n

7. DATA ACCESS AND USE

CC-BY 4.0\n\n###2. COMPETITION-SPECIFIC RULES \nIn addition to the provisions of the General Competition Rules below, you understand and agree to these Competition-Specific Rules required by the Competition Sponsor:\n\n####1. TEAM LIMITS\na. The maximum Team size is twenty (20).\nb. Team mergers are allowed and can be performed by the Team leader. In order to merge, the combined Team must have a total Submission count less than or equal to the maximum allowed as of the Team Merger Deadline. The maximum allowed is the number of Submissions per day multiplied by the number of days the competition has been running. For Hackathons, each team is allowed one (1) Submission; any Submissions submitted by Participants before merging into a Team will be unsubmitted.\n\n####2. SUBMISSION LIMITS\na. You may submit a maximum of one (1) Submissions per day.\nb. You may select up to one (1) Final Submissions for judging.\nc. For Hackathons, each Team may submit one (1) Submission only.\n\n####3. COMPETITION TIMELINE\na. Competition Timeline dates (including Entry Deadline, Final Submission Deadline, Start Date, and Team Merger Deadline, as applicable) are reflected on the competition’s Overview > Timeline page.\n\n#### 4. COMPETITION DATA\n\na. Data Access and Use. \n\n1. You may access and use the Competition Data for any purpose, whether commercial or non-commercial, including for participating in the Competition and on Kaggle.com forums, and for academic research and education. The Competition Sponsor reserves the right to disqualify any Participant who uses the Competition Data other than as permitted by the Competition Website and these Rules. \n\nb. Data Security.\n\n1. You agree to use reasonable and suitable measures to prevent persons who have not formally agreed to these Rules from gaining access to the Competition Data. You agree not to transmit, duplicate, publish, redistribute or otherwise provide or make available the Competition Data to any party not participating in the Competition. You agree to notify Kaggle immediately upon learning of any possible unauthorized transmission of or unauthorized access to the Competition Data and agree to work with Kaggle to rectify any unauthorized transmission or access.\n\n####5. WINNER LICENSE \n\na. Under Section 2.8 (Winners Obligations) of the General Rules below, you hereby grant and will grant the Competition Sponsor the following license(s) with respect to your Submission if you are a Competition winner:\n\n1. Open Source: You hereby license and will license your winning Submission and the source code used to generate the Submission under an Open Source Initiative-approved license (see [www.opensource.org] (http://www.opensource.org)) that in no event limits commercial use of such code or model containing or depending on such code. \n\n2. For generally commercially available software that you used to generate your Submission that is not owned by you, but that can be procured by the Competition Sponsor without undue expense, you do not need to grant the license in the preceding Section for that software. \n\n3. In the event that input data or pre-trained models with an incompatible license are used to generate your winning solution, you do not need to grant an open source license in the preceding Section for that data and/or model(s). \n\nb. You may be required by the Sponsor to provide a detailed description of how the winning Submission was generated, to the Competition Sponsor’s specifications, as outlined in Section 2.8, Winner’s Obligations. This may include a detailed description of methodology, where one must be able to reproduce the approach by reading the description, and includes a detailed explanation of the architecture, preprocessing, loss function, training details, hyper-parameters, etc. The description should also include a link to a code repository with complete and detailed instructions so that the results obtained can be reproduced.\n \n####6. EXTERNAL DATA AND TOOLS\n\na. You may use data other than the Competition Data (“External Data”) to develop and test your Submissions. However, you will ensure the External Data is either publicly available and equally accessible to use by all Participants of the Competition for purposes of the competition at no cost to the other Participants, or satisfies the Reasonableness criteria as outlined in Section 2.6.b below. The ability to use External Data under this Section does not limit your other obligations under these Competition Rules, including but not limited to Section 2.8 (Winners Obligations). \n\nb. The use of external data and models is acceptable unless specifically prohibited by the Host. Because of the potential costs or restrictions (e.g., “geo restrictions”) associated with obtaining rights to use external data or certain software and associated tools, their use must be “reasonably accessible to all” and of “minimal cost”. Also, regardless of the cost challenges as they might affect all Participants during the course of the competition, the costs of potentially procuring a license for software used to generate a Submission, must also be considered. The Host will employ an assessment of whether or not the following criteria can exclude the use of the particular LLM, data set(s), or tool(s):\n\n1. Are Participants being excluded from a competition because of the \"excessive\" costs for access to certain LLMs, external data, or tools that might be used by other Participants. The Host will assess the excessive cost concern by applying a “Reasonableness” standard (the “Reasonableness Standard”). The Reasonableness Standard will be determined and applied by the Host in light of things like cost thresholds and accessibility.\n\n2. By way of example only, a small subscription charge to use additional elements of a large language model such as Gemini Advanced are acceptable if meeting the Reasonableness Standard of Sec. 8.2. Purchasing a license to use a proprietary dataset that exceeds the cost of a prize in the competition would not be considered reasonable.\n\nc. Automated Machine Learning Tools (“AMLT”)\n\nWe consider two types of tools, subject to distinct rules, outlined below: \n\n1) “Automated Machine Learning Tools” (“AMLTs”), which denotes any prior software, not devised by the participant or Team, that essentially makes use of statistics to perform its function (e.g., all machine learning models, in particular, LLMs such as GPT-4, Claude, Gemini, Mistral, DeepSeekMath, LLaMA, as well as hybrid systems, such as AlphaGeometry).\n\n2) “Automated Deterministic Tools” (“ADTs”): which denote any other prior software, not devised by the participant or Team (e.g., Lean, Python runtimes used to execute code that an LLM may produce, etc.).\n\nThe participant or Team may either directly use any of these tools at runtime when the code is run to generate the output (direct modifications of AMLTs, such as fine-tuning an LLM, count as that AMLT being used at runtime), or may indirectly use any of these tools at some intermediate step in the pipeline or process necessary to generate the Submission (e.g., using GPT4 to generate data, which is subsequently used to fine-tune a different open-weight model that makes up the Submission), under the following conditions:\n\n- The way tools are incorporated must be compatible with the open source license of the final model (e.g., if you use a commercial tool that generates an output at some intermediate stage, which is further processed, you must ensure that you retain appropriate rights for that output)\n- If a submission makes essential use of an AMLT at runtime (e.g., using a finetuned version of an existing LLM on data generated by the participant or Team), we mandate that the AMLT be at least open-weight (e.g. DeepSeek-R1 may be used within a pipeline that outputs solutions to the problems, but GPT-5 may not)\n- All AMLTs you want to use at runtime need to be released prior to 15th March 2026. No time restriction is specified for AMLTs that are not used at runtime (e.g., you can use commercial LLMs to generate data that are released beyond October 1st).\n- Spot checks on submitted models will be conducted to assess whether the AMLT comply with the rules.\n- You are allowed to modify any tools you use (e.g., fine-tune an open-weight model) as long as these modifications are documented (see Section A.3) and fall within the specifications outlined by this document. In the case of ADTs, participants or Teams may only use open-source tools with appropriate licenses.\n\nThe essential criterion the Submission model must satisfy is that both running the model, as well as the process of arriving at the model must both be fully reproducible: This reproduction must be possible without undue costs and without access restrictions (in case commercial platforms were involved at intermediate steps towards the final Submission model).\n\n####7. ELIGIBILITY\n\na. Unless otherwise stated in the Competition-Specific Rules above or prohibited by internal policies of the Competition Entities, employees, interns, contractors, officers and directors of Competition Entities may enter and participate in the Competition, but are not eligible to win any Prizes. \"Competition Entities\" means the Competition Sponsor, Kaggle Inc., and their respective parent companies, subsidiaries and affiliates. If you are such a Participant from a Competition Entity, you are subject to all applicable internal policies of your employer with respect to your participation.\n\n####8. WINNER’S OBLIGATIONS\n\na. As a condition to being awarded a Prize, a Prize winner must fulfill the following obligations:\n\n1. Deliver to the Competition Sponsor the final model's software code as used to generate the winning Submission, as well as any artifacts (including, but not limited to, data, weights, and tools) used in the pipeline or process of arriving at the model’s software code. Running the Submission, as well as the process of arriving at the model, must be fully reproducible and clearly documented (using an automatic pipeline, Tensorflow TLX, to build the Submissions is preferred but not necessary). \n\n2. Supply the Competition Sponsor with all information he may request in relation to the model and its composition, such as training data (or the procedure used to construct the training data), training script, and the architecture and corresponding weights of the model. The delivered software code should follow [these Kaggle documentation guidelines](https://www.kaggle.com/WinningModelDocumentationGuidelines) and must be capable of generating the winning Submission by running, as a Jupyter Notebook, on the computing instance offered by Kaggle without any change required compared to the code at submission.\n\n3. If your winning Submission uses software or data that are subject to a license (e.g., open-weight LLMs, code libraries, datasets), you must provide a list of each such software or data used, as well as its corresponding license (see Section A.3). The description should also include a link to a code repository with complete and detailed instructions so that the results obtained can be reproduced without undue effort.\n\n4. To the extent that the final model’s software code includes generally commercially available software that is not owned by you, but that can be procured by the Competition Sponsor without undue expense, then instead of delivering the code for that software to the Competition Sponsor, you must identify that software, method for procuring it, and any parameters or other information necessary to replicate the winning Submission; Individual Participants and Teams who create a Submission using an AMLT may win a Prize. However, for clarity, the potential winner’s Submission must still meet the requirements of these Rules, including but not limited to Section 2.5 (Winners License), Section 2.8 (Winners Obligations), and Section 3.14 (Warranty, Indemnity, and Release).” \n\n5. Individual Participants and Teams who create a Submission using an AMLT may win a Prize. However, for clarity, the potential winner’s Submission must still meet the requirements of these Rules.\n\nb. In addition to the previously mentioned documentation guidelines, the participant or Team also provides documentation that contains a description of computational resources required to build and/or run the final Submission of executable code successfully, as well as detailed documentation consisting of (but not limited to) around the design choices of the code, and potential dataset creation. To achieve this, the host will send you a questionnaire for completion. \n\nc. *Hosting and Dissemination*: The participant or Team agrees to host the final model, on which his/her final Submission is based on, unchanged, on HuggingFace, enabling the model to be run on one of the deployment services (e.g., dedicated Inference Endpoint). Additionally, you need to follow our instructions to containerize your solution in a [Singularity container](https://sylabs.io/singularity/), based on the interface you will be provided in the questionnaire.\n\nd. You further agree to work with the Hosts to remove Kaggle-related compute restrictions of your model (necessary to ensure output given the runtime). This improved version also will be hosted on HuggingFace and provided as a Singularity container to the Hosts.\n\ne. You agree to work with the Hosts to run internal benchmarks on both the original winning model, as well as any improved version of your model, and to participate in the writing of a paper, directed by the Hosts, that showcases the Competition's winning models.\n\nf. *License*: The participant or Team agrees to grant to the Competition Sponsor the license to the winning Submission stated in the Competition Specific Rules above, and represent that they have the unrestricted right to grant that license.\n\ng. *AML/KYC checks*: sign and return all Prize acceptance documents as may be required by Competition Sponsor or Kaggle and provide all information the Competition Sponsor may require to meet its AML/KYC checks, including without limitation: (i) eligibility certifications; (ii) identity documentation such as passport or driving license and proof of address; (iii) licenses, releases and other agreements required under the Rules; and (iv) U.S. tax forms (such as IRS Form W-9 if U.S. resident, IRS Form W-8BEN if foreign resident, or future equivalents).\n\nh. Grant to the Competition Sponsor the license to the winning Submission stated in the Competition Specific Rules above, and represent that you have the unrestricted right to grant that license; \n\ni. Sign and return all Prize acceptance documents as may be required by Competition Sponsor or Kaggle, including without limitation: (a) eligibility certifications; (b) licenses, releases and other agreements required under the Rules; and (c) U.S. tax forms (such as IRS Form W-9 if U.S. resident, IRS Form W-8BEN if foreign resident, or future equivalents).\n\n####9. GOVERNING LAW\n\na. Unless otherwise provided in the Competition Specific Rules above, all claims arising out of or relating to these Rules (including any non-contractual obligations) will be governed by English law and will be litigated exclusively in English courts. The parties consent to personal jurisdiction in those courts. If any provision of these Rules is held to be invalid or unenforceable, all remaining provisions of the Rules will remain in full force and effect.\n\n###3. GENERAL COMPETITION RULES - BINDING AGREEMENT\n####1. ELIGIBILITY\na. To be eligible to enter the Competition, you must be:\n1. a registered account holder at Kaggle.com; \n2. the older of 18 years old or the age of majority in your jurisdiction of residence (unless otherwise agreed to by Competition Sponsor and appropriate parental/guardian consents have been obtained by Competition Sponsor); \n3. not a resident of Crimea, so-called Donetsk People's Republic (DNR) or Luhansk People's Republic (LNR), Cuba, Iran, or North Korea; and\n4. not a person or representative of an entity under U.S. export controls or sanctions (see: [https://www.treasury.gov/resourcecenter/sanctions/Programs/Pages/Programs.aspx][1]).\n\nb. Competitions are open to residents of the United States and worldwide, except that if you are a resident of Crimea, so-called Donetsk People's Republic (DNR) or Luhansk People's Republic (LNR), Cuba, Iran, North Korea, or are subject to U.S. export controls or sanctions, you may not enter the Competition. Other local rules and regulations may apply to you, so please check your local laws to ensure that you are eligible to participate in skills-based competitions. The Competition Host reserves the right to forego or award alternative Prizes where needed to comply with local laws. If a winner is located in a country where prizes cannot be awarded, then they are not eligible to receive a prize.\n\nc. If you are entering as a representative of a company, educational institution or other legal entity, or on behalf of your employer, these rules are binding on you, individually, and the entity you represent or where you are an employee. If you are acting within the scope of your employment, or as an agent of another party, you warrant that such party or your employer has full knowledge of your actions and has consented thereto, including your potential receipt of a Prize. You further warrant that your actions do not violate your employer's or entity's policies and procedures. \n\nd. The Competition Sponsor reserves the right to verify eligibility and to adjudicate on any dispute at any time. If you provide any false information relating to the Competition concerning your identity, residency, mailing address, telephone number, email address, ownership of right, or information required for entering the Competition, you may be immediately disqualified from the Competition.\n\n####2. SPONSOR AND HOSTING PLATFORM\n\na. The Competition is sponsored by Competition Sponsor named above. The Competition is hosted on behalf of Competition Sponsor by Kaggle Inc. (\"Kaggle\"). Kaggle is an independent contractor of Competition Sponsor, and is not a party to this or any agreement between you and Competition Sponsor. You understand that Kaggle has no responsibility with respect to selecting the potential Competition winner(s) or awarding any Prizes. Kaggle will perform certain administrative functions relating to hosting the Competition, and you agree to abide by the provisions relating to Kaggle under these Rules. As a Kaggle.com account holder and user of the Kaggle competition platform, remember you have accepted and are subject to the Kaggle Terms of Service at [www.kaggle.com/terms][2] in addition to these Rules.\n\n####3. COMPETITION PERIOD\na. For the purposes of Prizes, the Competition will run from the Start Date and time to the Final Submission Deadline (such duration the “Competition Period”). The Competition Timeline is subject to change, and Competition Sponsor may introduce additional hurdle deadlines during the Competition Period. Any updated or additional deadlines will be publicized on the Competition Website. It is your responsibility to check the Competition Website regularly to stay informed of any deadline changes. YOU ARE RESPONSIBLE FOR DETERMINING THE CORRESPONDING TIME ZONE IN YOUR LOCATION.\n\n####4. COMPETITION ENTRY\na. NO PURCHASE NECESSARY TO ENTER OR WIN. To enter the Competition, you must register on the Competition Website prior to the Entry Deadline, and follow the instructions for developing and entering your Submission through the Competition Website. Your Submissions must be made in the manner and format, and in compliance with all other requirements, stated on the Competition Website (the \"Requirements\"). Submissions must be received before any Submission deadlines stated on the Competition Website. Submissions not received by the stated deadlines will not be eligible to receive a Prize.\nb. Except as expressly allowed in Hackathons as set forth on the Competition Website, submissions may not use or incorporate information from hand labeling or human prediction of the validation dataset or test data records.\nc. If the Competition is a multi-stage competition with temporally separate training and/or test data, one or more valid Submissions may be required during each Competition stage in the manner described on the Competition Website in order for the Submissions to be Prize eligible.\nd. Submissions are void if they are in whole or part illegible, incomplete, damaged, altered, counterfeit, obtained through fraud, or late. Competition Sponsor reserves the right to disqualify any entrant who does not follow these Rules, including making a Submission that does not meet the Requirements. \n\n####5. INDIVIDUALS AND TEAMS\na. Individual Account. You may make Submissions only under one, unique Kaggle.com account. You will be disqualified if you make Submissions through more than one Kaggle account, or attempt to falsify an account to act as your proxy. You may submit up to the maximum number of Submissions per day as specified on the Competition Website. \nb. Teams. If permitted under the Competition Website guidelines, multiple individuals may collaborate as a Team; however, you may join or form only one Team. Each Team member must be a single individual with a separate Kaggle account. You must register individually for the Competition before joining a Team. You must confirm your Team membership to make it official by responding to the Team notification message sent to your Kaggle account. Team membership may not exceed the Maximum Team Size stated on the Competition Website.\nc. Team Merger. Teams (or individual Participants) may request to merge via the Competition Website. Team mergers may be allowed provided that: (i) the combined Team does not exceed the Maximum Team Size; (ii) the number of Submissions made by the merging Teams does not exceed the number of Submissions permissible for one Team at the date of the merger request; (iii) the merger is completed before the earlier of: any merger deadline or the Competition deadline; and (iv) the proposed combined Team otherwise meets all the requirements of these Rules. \nd. Private Sharing. No private sharing outside of Teams. Privately sharing code or data outside of Teams is not permitted. It's okay to share code if made available to all Participants on the forums.\n\n####6. SUBMISSION CODE REQUIREMENTS\na. Private Code Sharing. Unless otherwise specifically permitted under the Competition Website or Competition Specific Rules above, during the Competition Period, you are not allowed to privately share source or executable code developed in connection with or based upon the Competition Data or other source or executable code relevant to the Competition (“Competition Code”). This prohibition includes sharing Competition Code between separate Teams, unless a Team merger occurs. Any such sharing of Competition Code is a breach of these Competition Rules and may result in disqualification.\nb. Public Code Sharing. You are permitted to publicly share Competition Code, provided that such public sharing does not violate the intellectual property rights of any third party. If you do choose to share Competition Code or other such code, you are required to share it on Kaggle.com on the discussion forum or notebooks associated specifically with the Competition for the benefit of all competitors. By so sharing, you are deemed to have licensed the shared code under an Open Source Initiative-approved license (see [www.opensource.org][3]) that in no event limits commercial use of such Competition Code or model containing or depending on such Competition Code.\nc. Use of Open Source. Unless otherwise stated in the Specific Competition Rules above, if open source code is used in the model to generate the Submission, then you must only use open source code licensed under an Open Source Initiative-approved license (see [www.opensource.org][4]) that in no event limits commercial use of such code or model containing or depending on such code.\n\n####7. DETERMINING WINNERS\na. Each Submission will be scored and/or ranked by the evaluation metric, or Evaluation Rubric (in the case of Hackathon Competitions),stated on the Competition Website. During the Competition Period, the current ranking will be visible on the Competition Website's Public Leaderboard. The potential winner(s) are determined solely by the leaderboard ranking on the Private Leaderboard, subject to compliance with these Rules. The Public Leaderboard will be based on the public test set and the Private Leaderboard will be based on the private test set. There will be no leaderboards for Hackathon Competitions.\nb. In the event of a tie, the Submission that was entered first to the Competition will be the winner. In the event a potential winner is disqualified for any reason, the Submission that received the next highest score rank will be chosen as the potential winner. For Hackathon Competitions, each of the top Submissions will get a unique ranking and there will be no tiebreakers. \n\n####8. NOTIFICATION OF WINNERS & DISQUALIFICATION\na. The potential winner(s) will be notified by email. \nb. If a potential winner (i) does not respond to the notification attempt within one (1) week from the first notification attempt or (ii) notifies Kaggle within one week after the Final Submission Deadline that the potential winner does not want to be nominated as a winner or does not want to receive a Prize, then, in each case (i) and (ii) such potential winner will not receive any Prize, and an alternate potential winner will be selected from among all eligible entries received based on the Competition’s judging criteria.\nc. In case (i) and (ii) above Kaggle may disqualify the Participant. However, in case (ii) above, if requested by Kaggle, such potential winner may provide code and documentation to verify the Participant’s compliance with these Rules. If the potential winner provides code and documentation to the satisfaction of Kaggle, the Participant will not be disqualified pursuant to this paragraph.\nd. Competition Sponsor reserves the right to disqualify any Participant from the Competition if the Competition Sponsor reasonably believes that the Participant has attempted to undermine the legitimate operation of the Competition by cheating, deception, or other unfair playing practices or abuses, threatens or harasses any other Participants, Competition Sponsor or Kaggle.\ne. A disqualified Participant may be removed from the Competition leaderboard, at Kaggle's sole discretion. If a Participant is removed from the Competition Leaderboard, additional winning features associated with the Kaggle competition platform, for example Kaggle points or medals, may also not be awarded.\nf. The final leaderboard list will be publicly displayed at Kaggle.com. Determinations of Competition Sponsor are final and binding.\n\n####9. PRIZES\na. Prize(s) are as described on the Competition Website and are only available for winning during the time period described on the Competition Website. The odds of winning any Prize depends on the number of eligible Submissions received during the Competition Period and the skill of the Participants. \nb. All Prizes are subject to Competition Sponsor's review and verification of the Participant’s eligibility and compliance with these Rules, and the compliance of the winning Submissions with the Submissions Requirements. In the event that the Submission demonstrates non-compliance with these Competition Rules, Competition Sponsor may at its discretion take either of the following actions: (i) disqualify the Submission(s); or (ii) require the potential winner to remediate within one week after notice all issues identified in the Submission(s) (including, without limitation, the resolution of license conflicts, the fulfillment of all obligations required by software licenses, and the removal of any software that violates the software restrictions).\nc. A potential winner may decline to be nominated as a Competition winner in accordance with Section 3.8.\nd. Potential winners must return all required Prize acceptance documents within two (2) weeks following notification of such required documents, or such potential winner will be deemed to have forfeited the prize and another potential winner will be selected. Prize(s) will be awarded within approximately thirty (30) days after receipt by Competition Sponsor or Kaggle of the required Prize acceptance documents. Transfer or assignment of a Prize is not allowed. \ne. You are not eligible to receive any Prize if you do not meet the Eligibility requirements in Section 2.7 and Section 3.1 above.\nf. If a Team wins a monetary Prize, the Prize money will be allocated in even shares between the eligible Team members, unless the Team unanimously opts for a different Prize split and notifies Kaggle before Prizes are issued.\n\n####10. TAXES\na. ALL TAXES IMPOSED ON PRIZES ARE THE SOLE RESPONSIBILITY OF THE WINNERS. Payments to potential winners are subject to the express requirement that they submit all documentation requested by Competition Sponsor or Kaggle for compliance with applicable state, federal, local and foreign (including provincial) tax reporting and withholding requirements. Prizes will be net of any taxes that Competition Sponsor is required by law to withhold. If a potential winner fails to provide any required documentation or comply with applicable laws, the Prize may be forfeited and Competition Sponsor may select an alternative potential winner. Any winners who are U.S. residents will receive an IRS Form-1099 in the amount of their Prize.\n\n####11. GENERAL CONDITIONS\na. All federal, state, provincial and local laws and regulations apply.\n\n####12. PUBLICITY\na. You agree that Competition Sponsor, Kaggle and its affiliates may use your name and likeness for advertising and promotional purposes without additional compensation, unless prohibited by law.\n\n####13. PRIVACY\na. You acknowledge and agree that Competition Sponsor and Kaggle may collect, store, share and otherwise use personally identifiable information provided by you during the Kaggle account registration process and the Competition, including but not limited to, name, mailing address, phone number, and email address (“Personal Information”). Kaggle acts as an independent controller with regard to its collection, storage, sharing, and other use of this Personal Information, and will use this Personal Information in accordance with its Privacy Policy <[www.kaggle.com/privacy][6]>, including for administering the Competition. As a Kaggle.com account holder, you have the right to request access to, review, rectification, portability or deletion of any personal data held by Kaggle about you by logging into your account and/or contacting Kaggle Support at <[www.kaggle.com/contact][7]>.\nb. As part of Competition Sponsor performing this contract between you and the Competition Sponsor, Kaggle will transfer your Personal Information to Competition Sponsor, which acts as an independent controller with regard to this Personal Information. As a controller of such Personal Information, Competition Sponsor agrees to comply with all U.S. and foreign data protection obligations with regard to your Personal Information. Kaggle will transfer your Personal Information to Competition Sponsor in the country specified in the Competition Sponsor Address listed above, which may be a country outside the country of your residence. Such country may not have privacy laws and regulations similar to those of the country of your residence.\n\n####14. WARRANTY, INDEMNITY AND RELEASE \na. You warrant that your Submission is your own original work and, as such, you are the sole and exclusive owner and rights holder of the Submission, and you have the right to make the Submission and grant all required licenses. You agree not to make any Submission that: (i) infringes any third party proprietary rights, intellectual property rights, industrial property rights, personal or moral rights or any other rights, including without limitation, copyright, trademark, patent, trade secret, privacy, publicity or confidentiality obligations, or defames any person; or (ii) otherwise violates any applicable U.S. or foreign state or federal law.\nb. To the maximum extent permitted by law, you indemnify and agree to keep indemnified Competition Entities at all times from and against any liability, claims, demands, losses, damages, costs and expenses resulting from any of your acts, defaults or omissions and/or a breach of any warranty set forth herein. To the maximum extent permitted by law, you agree to defend, indemnify and hold harmless the Competition Entities from and against any and all claims, actions, suits or proceedings, as well as any and all losses, liabilities, damages, costs and expenses (including reasonable attorneys fees) arising out of or accruing from: (a) your Submission or other material uploaded or otherwise provided by you that infringes any third party proprietary rights, intellectual property rights, industrial property rights, personal or moral rights or any other rights, including without limitation, copyright, trademark, patent, trade secret, privacy, publicity or confidentiality obligations, or defames any person; (b) any misrepresentation made by you in connection with the Competition; (c) any non-compliance by you with these Rules or any applicable U.S. or foreign state or federal law; (d) claims brought by persons or entities other than the parties to these Rules arising from or related to your involvement with the Competition; and (e) your acceptance, possession, misuse or use of any Prize, or your participation in the Competition and any Competition-related activity.\nc. You hereby release Competition Entities from any liability associated with: (a) any malfunction or other problem with the Competition Website; (b) any error in the collection, processing, or retention of any Submission; or (c) any typographical or other error in the printing, offering or announcement of any Prize or winners.\n\n####15. INTERNET\na. Competition Entities are not responsible for any malfunction of the Competition Website or any late, lost, damaged, misdirected, incomplete, illegible, undeliverable, or destroyed Submissions or entry materials due to system errors, failed, incomplete or garbled computer or other telecommunication transmission malfunctions, hardware or software failures of any kind, lost or unavailable network connections, typographical or system/human errors and failures, technical malfunction(s) of any telephone network or lines, cable connections, satellite transmissions, servers or providers, or computer equipment, traffic congestion on the Internet or at the Competition Website, or any combination thereof, which may limit a Participant’s ability to participate.\n\n####16. RIGHT TO CANCEL, MODIFY OR DISQUALIFY\na. If for any reason the Competition is not capable of running as planned, including infection by computer virus, bugs, tampering, unauthorized intervention, fraud, technical failures, or any other causes which corrupt or affect the administration, security, fairness, integrity, or proper conduct of the Competition, Competition Sponsor reserves the right to cancel, terminate, modify or suspend the Competition. Competition Sponsor further reserves the right to disqualify any Participant who tampers with the submission process or any other part of the Competition or Competition Website. Any attempt by a Participant to deliberately damage any website, including the Competition Website, or undermine the legitimate operation of the Competition is a violation of criminal and civil laws. Should such an attempt be made, Competition Sponsor and Kaggle each reserves the right to seek damages from any such Participant to the fullest extent of the applicable law.\n\n####17. NOT AN OFFER OR CONTRACT OF EMPLOYMENT\na. Under no circumstances will the entry of a Submission, the awarding of a Prize, or anything in these Rules be construed as an offer or contract of employment with Competition Sponsor or any of the Competition Entities. You acknowledge that you have submitted your Submission voluntarily and not in confidence or in trust. You acknowledge that no confidential, fiduciary, agency, employment or other similar relationship is created between you and Competition Sponsor or any of the Competition Entities by your acceptance of these Rules or your entry of your Submission.\n\n####18. DEFINITIONS\na. \"Competition Data\" are the data or datasets available from the Competition Website for the purpose of use in the Competition, including any prototype or executable code provided on the Competition Website. The Competition Data will contain private and public test sets. Which data belongs to which set will not be made available to Participants. \nb. An “Entry” is when a Participant has joined, signed up, or accepted the rules of a competition. Entry is required to make a Submission to a competition.\nc. A “Final Submission” is the Submission selected by the user, or automatically selected by Kaggle in the event not selected by the user, that is/are used for final placement on the competition leaderboard.\nd. A “Participant” or “Participant User” is an individual who participates in a competition by entering the competition and making a Submission.\ne. The “Private Leaderboard” is a ranked display of Participants’ Submission scores against the private test set. The Private Leaderboard determines the final standing in the competition.\nf. The “Public Leaderboard” is a ranked display of Participants’ Submission scores against a representative sample of the test data. This leaderboard is visible throughout the competition.\ng. A “Sponsor” is responsible for hosting the competition, which includes but is not limited to providing the data for the competition, determining winners, and enforcing competition rules.\nh. A “Submission” is anything provided by the Participant to the Sponsor to be evaluated for competition purposes and determine leaderboard position. A Submission may be made as a model, notebook, prediction file, or other format as determined by the Sponsor.\ni. A “Team” is one or more Participants participating together in a Kaggle competition, by officially merging together as a Team within the competition platform.\n\n [1]: https://www.treasury.gov/resource-center/sanctions/Programs/Pages/Programs.aspx\n [2]: http://www.kaggle.com/terms\n [3]: http://www.opensource.org\n [4]: http://www.opensource.org\n [5]: https://www.kaggle.com/WinningModelDocumentationGuidelines\n [6]: http://www.kaggle.com/privacy\n [7]: http://www.kaggle.com/contact\n\n\n\n\n\n\n\n\n\n\n\n", "isPublished": false, "hasIsPublished": false, "postTitle": "", "hasPostTitle": false, "mimeType": "", "hasMimeType": false }, { "name": "Description", "content": "*Note: This is the third [AIMO Progress Prize](https://aimoprize.com/) competition. It builds upon the [first AIMO Progress Prize competition](https://www.kaggle.com/c/ai-mathematical-olympiad-prize), which was won in July 2024 by Project Numina and the [second AIMO Progress Prize competition](https://www.kaggle.com/competitions/ai-mathematical-olympiad-progress-prize-2), which was won in April 2025 by Nvidia’s NemoSkills team. This third competition features a significantly increased problem difficulty, new submission format, expanded prize pool, industry-leading compute resources for participants, new auxiliary prizes to reward community contributions, and updated rules for using open-source LLMs.*\n\nThe ability to reason mathematically is a critical milestone for AI. Mathematical reasoning is the foundation for solving many complex problems, from engineering marvels to intricate financial models. However, current AI capabilities are limited in this area.\n\nThe AI Mathematical Olympiad (AIMO) Prize is a $10mn fund to spur the open development of AI models capable of performing as well as top human participants in the International Mathematical Olympiad (IMO).\n\nRecent breakthroughs demonstrate AI achieving human-level results on IMO problems. Closed-source models achieved gold medal performance at the 2025 IMO, demonstrating that frontier AI systems are now capable of solving the world's most challenging high school mathematics problems. Our [\"OpenAI x AIMO eval\"](https://aimoprize.com/updates/2025-09-05-the-gap-is-shrinking) in March 2025 revealed that commercial models could solve 50/50 AIMO2 public leaderboard problems with sufficient compute—despite the highest Kaggle score being just 34/50. This demonstrates a significant gap between closed-source and open-source models in mathematical reasoning capabilities.\n\nAIMO3 is designed to accelerate open-source progress and close this gap. The competition features 110 problems spanning algebra, combinatorics, geometry, and number theory, ranging from National Olympiad level all the way up to IMO standard—the pinnacle of high school mathematical achievement. All problems are entirely original, created by an international team of problem solvers to ensure zero risk of train-test contamination. They have been carefully designed to be 'AI hard', tested against current open LLMs' capabilities. They require genuine mathematical reasoning to solve, making answer-only testing as robust as possible (see Problem 10 in the Reference Problems PDF for a detailed discussion of this design philosophy).\n\nEvery problem is entirely original, ensuring zero risk of data contamination. Using this transparent and fair evaluation framework, the competition will help to strengthen the benchmarks for assessing AI models' mathematical reasoning skills. This latest AIMO Progress Prize competition offers an exciting opportunity to drive innovation in the field of AI for Math, while also fostering healthy competition and supporting open science.\n\nJoin us as we work towards a future where AI models’ mathematical reasoning skills are accurately and reliably assessed, driving progress and innovation.\n\n", "isPublished": false, "hasIsPublished": false, "postTitle": "", "hasPostTitle": false, "mimeType": "", "hasMimeType": false }, { "name": "Evaluation", "content": "During the submission period, submission notebooks are run only over the public test set and are evaluated by the unnormalized [**accuracy**](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.accuracy_score.html) between their predicted labels and the ground-truth labels (i.e. number of correct answers).\n\nAfter the submission deadline, submission notebooks will be run *twice* over the private test set and their predictions concatenated into a single submission file. We then evaluate submissions by a **penalized accuracy** score, as follows:\n- If both predicted answers for a problem are correct, the score is 1 for that problem.\n- If one predicted answer is correct and the other is incorrect, the score is 0.5 for that problem.\n- If neither predicted answer is correct, the score is 0 for that problem.\n \nA submission's overall score is the sum of its scores for each problem.\n\nIn this competition, every ground-truth label is an integer between 0 and 99999, inclusive. Any modulo calculation required is explicitly stated in the problem statement (e.g., \"What is the remainder when \\\\(N\\\\) is divided by \\\\(10^5\\\\)?\"), meaning all problems have answers in this range without any further adjustments.\n\n**Note:** This is a change from the first and second Progress Prizes, which used 3-digit answers with an implicit requirement to take your answer modulo 1000.\n\nAnswers may require basic computations, including modular arithmetic. For example, \\\\(\\left\\lfloor 10^4\\sqrt{2}\\right\\rfloor=14142\\\\) and \\\\(3^{2025} \\bmod{10^5} = 29443\\\\). Additional examples of computations that may be required are provided in the Reference Problems PDF found on the [Data](https://www.kaggle.com/competitions/ai-mathematical-olympiad-progress-prize-3/data) page.\n\n## Submitting ##\n\nYou must submit to this competition using the provided Python evaluation API, which serves test set instances one-by-one in random order for the public leaderboard, and uses fixed, random order for the private leaderboard. To use the API, follow the template in [this notebook](https://www.kaggle.com/code/ryanholbrook/aimo-3-submission-demo).\n", "isPublished": false, "hasIsPublished": false, "postTitle": "", "hasPostTitle": false, "mimeType": "", "hasMimeType": false }, { "name": "Prizes", "content": "##**TOTAL FUND FOR PROGRESS PRIZE 3: $2,207,152**\n\n###**Prizes for Top-Ranking Teams in this Competition**:\n- 1st Place: $262,144\n- 2nd Place: $131,072\n- 3rd Place: $65,536\n- 4th Place: $32,768\n- 5th Place: $16,384\n\n**Overall Progress Prize Winner**: The Overall Progress Prize Winner shall be the highest ranking team that achieves a score of at least 47/50 on both public and private test sets. After any prizes for the five top-ranking teams have been awarded, the remainder of the total fund shall be awarded to the Overall Progress Prize Winner.\n\nIf a team is named the Overall Progress Prize Winner in this competition, the prize will be at least $1,589,248. If no team is named the Overall Progress Prize Winner in this competition, the remainder of the total fund shall roll over to the next competition, where the same prize allocation will apply.\n\n### Additional Prizes: $110,000\n\n- **Longest Leader Prize**: $20,000\n- **Hard Problem Prize**: $30,000\n- **Math Corpus Prize**: $30,000\n- **Writeup Prizes**: 2x $15,000 \n\n####Longest Leader Prize: $20,000\nAwarded to the team whose notebook(s) generates the best scoring submission on the leaderboard for the longest period of time between November 20, 2025 and February 2, 2026 11:59 PM UTC. The notebooks need to adhere to the same requirements and restrictions regarding licensing, reproducibility, and documentation to which the winning Submission is subject (see [Competition Rules](https://www.kaggle.com/competitions/ai-mathematical-olympiad-progress-prize-3/rules)). The notebook(s) and any datasets used must be made publicly available at latest at February 9, 2026 11:59 PM UTC and kept public until the final Progress Prizes are awarded to the winning Teams at the end of the competition. In the event that a team has multiple public notebooks, each achieving the top leaderboard position at different times, those durations will be combined.\n\n\n####Hard Problem Prize: $30,000\nAwarded to the highest ranked team(s) on the private leaderboard who solved the most difficult problem(s) in the private test set. Problem difficulty is measured by the average accuracy score across all selected submissions at the end of the competition, where average accuracy is calculated as described in the Evaluation section (accounting for both runs of each submission). For example, if the most difficult problem has an average accuracy of 1.7%, then the highest ranked team on the private leaderboard who achieved a full score on that problem will receive the prize. If multiple problems tie for \"most difficult\": We will identify the highest ranked team for each tied problem separately, and the $30,000 prize will be split equally among the winning teams. For example, if two problems tie for most difficult, each with its own highest-scoring solver, the prize will be divided into two $15,000 awards. In the event of a tie among notebooks, the tiebreaker will be based on submission time (first to submit).\n\n#### Math Corpus Prize: $30,000 \nAwarded to the team who provides the most valuable dataset to the community for improving AIMO model performance.\n\n*Math Corpus Prize Requirements*\n- The dataset must be publicly released either on Kaggle or on HuggingFace prior to February 9, 2026 11:59 PM UTC. Teams must create a Kaggle Discussion post until this date that explicitly links to their dataset and tags it as an entry for the Math Corpus Prize. All eligible datasets will be judged according to the criteria below, and the highest scoring dataset will be awarded the prize.\n\n- Your Kaggle Discussion post will be used to judge your dataset submission based on the criteria below, and your claims will be independently verified by the Host Team.\n- The dataset must be in English.\n- The earliest timestamped version of the dataset will be used for evaluation (later modifications will not be considered)\n- The dataset is not allowed to exceed 5M datapoints, which is an upper bound on common datasets used by prior AIMO winners to fine-tune their models. Each datapoint is not allowed to exceed more than 100k characters. \n- Each datapoint from the submitted dataset must come with an open source license that allows free dissemination of data.\n\n*Math Corpus Prize Evaluation Criteria*:\n\n- Data Novelty (25 points) \n - Is the dataset distinctly unique from other datasets found on the internet? Minor modification are not considered novel. In particular, translations into English from other languages are not considered significant modifications.\n - Describe what methods you used to ensure your dataset is novel, and note how your dataset does not overlap with prior submitted datasets.\n- Format (25 points) \n - Does the dataset comes in a format that makes it easy to handle and train a model?\n - Does each datapoint contain rich metadata?\n- Performance (50 points) \n - Does the dataset improve mathematical reasoning and aid in improving model performance during the AIMO3. How so and how much?\n\n#### Writeup Prize: $30,000 (2 awards x $15,000)\nAwarded to the two best Solution Writeups from non-winning teams based on the below criteria.\n\n*Writeup Prize Requirements*\n- The Writeup should be an official [Kaggle Writeup](https://www.kaggle.com/discussions/product-announcements/593763), attached to your submission within one week after the competition ended.\n- It should mirror an academic publication about your submission, and include graphs that outline how your model was trained, the impact of different techniques and methods on performance, and how it could be further improved. (The arXiv post by the AIMO2 winners is a good reference paper in terms of academic standard: https://arxiv.org/abs/2504.16891)\n- It should be limited to 30k words (which approximately corresponds to a 50-page paper), and split into a main section of not more than 5k words (approximately 10 pages), which outlines your methodology, and the remaining sections, which include the bibliography, examples, and any other materials that support the main section.\n\n*Writeup Prize Evaluation Criteria*\n- Clarity (10 points)\n - Does the Writeup describe in detail the entire model creation lifecycle (including data collection, training, and benchmarking), as well as any model variants you explored?\n- Ablation Studies and Variants (10 points)\n - Does the Writeup contain information that pinpoints where the performance gains of your model came from? \n- Comparison with the State of the Art (10 points)\n - Does the Writeup evaluate how your model (potentially split across types of reasoning modes or mathematical problems) compares to the state of the art for both open-weight and commercial models?\n- Graphs and Charts (10 points) \n - Does the Writeup contain well-designed visuals that clearly \"tell the story\" by illustrating your model's performance, training process, and key experimental results?\n- Reproducibility (60 points) \n - Is your write-up sufficiently detailed to easily allow the reader to reproduce the training of your model starting from the same artifacts (base open-weight models and data) that you started from? \n", "isPublished": false, "hasIsPublished": false, "postTitle": "", "hasPostTitle": false, "mimeType": "", "hasMimeType": false }, { "name": "data-description", "content": "The competition data comprises 110 mathematics problems similar in style to those of the [AIME][1].\n\nThe answer to each problem is a non-negative integer between `0` and `99999`, inclusive.\n\nYou should expect the problems to be roughly ranging in difficulty from a national Olympiad to the International Mathematical Olympiad (IMO), with some problems being slightly easier and some being slightly harder.\n\nAll problems are text-only with mathematical notation in LaTeX. Please see the Overview, Section [*Note on Language and Notation*](https://www.kaggle.com/c/ai-mathematical-olympiad-progress-prize-3/overview/note-on-language-and-notation) for details on the notational conventions used. \n\nAlthough some problems may involve geometry, diagrams are not used in any problem.\n\nThe public test set comprises exactly 50 problems, and the private test set comprises a further distinct set of 50 problems. We also provide a selection of 10 problems for use as reference, called \"*reference data*\". A pdf with full solutions to the problems from the reference data can be found [below](https://www.kaggle.com/c/ai-mathematical-olympiad-progress-prize-3/data?select=AIMO_Progress_Prize_3_Reference_Problems.pdf). The problems in the private and public sets have been selected to balance both difficulty and subject area. Their difficulty and subject distribution is similar.\n\n### Competition Data\n\nPlease note that this is a Code Competition. You must submit to this competition using the provided Python evaluation API, which serves the test set question by question **in random order**. To use the API, follow the example in this notebook: [**AIMO 3 Submission Demo**](https://www.kaggle.com/code/ryanholbrook/aimo-3-submission-demo).\n\nWe give a few placeholder problems in the publicly visible `test.csv` file to help you author your submissions. These problems are not representative of the problems in the actual test set.\n\nWhen your submission is scored, this placeholder test data will be replaced with the actual test data.\n\n### Private Set Rerun\n\nBecause of the limited number of problems available, we are taking special precautions to secure the test set against probing. Among other things, during the submission period, the test set will comprise only the 50 public set problems. Once the competition ends, we will rerun all model submissions *twice* on the 50 private set problems. The overall score is an average accuracy score as described on the [Evaluation](https://www.kaggle.com/c/ai-mathematical-olympiad-progress-prize-3/overview/evaluation) page. You should attempt to make sure your submission will complete successfully on the 50 new private set problems. This may mean ensuring your submission is robust to unexpected inputs, or managing runtime and memory usage. A submission must complete successfully on both private set reruns or it will not be scored.\n\n### Files\n\n- **reference.csv** - Contains 10 problems for use as reference data. \n- **test.csv** - Contains 50 problems. Please note that the problems visible here are only placeholders. Your submission will have access to the full set of problems during scoring. \n- **sample_submission.csv** - A sample submission file in the correct format. See the [Evaluation](https://www.kaggle.com/c/ai-mathematical-olympiad-progress-prize-3/overview/evaluation) page for more information on the submission format.\n\n### Fields\n\n- `id` - A unique identifier for each problem. \n- `problem` - A LaTeX statement of the problem to be solved. \n- `answer` - An integer from `0` to `99999`, inclusive.\n\n\n [1]: https://artofproblemsolving.com/wiki/index.php/American_Invitational_Mathematics_Examination", "isPublished": false, "hasIsPublished": false, "postTitle": "", "hasPostTitle": false, "mimeType": "", "hasMimeType": false }, { "name": "Timeline", "content": "* **November 20, 2025** - Start Date. \n* **April 8, 2026** - Entry Deadline. You must accept the competition rules before this date in order to compete. \n* **April 8, 2026** - Team Merger Deadline. This is the last day participants may join or merge teams. \n* **April 15, 2026** - Final Submission Deadline. \n\nAs noted in the [Evaluation](https://www.kaggle.com/competitions/ai-mathematical-olympiad-progress-prize-3/overview/evaluation) section, selected submissions will be run against the private test set after the Submission deadline. Final results will be available shortly thereafter.\n\nAll deadlines are at 11:59 PM UTC on the corresponding day unless otherwise noted. The competition organizers reserve the right to update the contest timeline if they deem it necessary.", "isPublished": false, "hasIsPublished": false, "postTitle": "", "hasPostTitle": false, "mimeType": "", "hasMimeType": false }, { "name": "abstract", "content": "The goal of this competition is to create open-source algorithms and models that can solve olympiad-level math problems written in LaTeX format. Your participation will help to advance AI models’ mathematical reasoning skills and drive frontier knowledge.", "isPublished": false, "hasIsPublished": false, "postTitle": "", "hasPostTitle": false, "mimeType": "", "hasMimeType": false }, { "name": "Code Requirements", "content": "\n\nSubmissions to this competition must be made through Notebooks. In order for the \"Submit\" button to be active after a commit, the following conditions must be met: \n- CPU Notebook <= 9 hours run-time \n- GPU Notebook <= 5 hours run-time \n- Internet access disabled \n- Freely & publicly available external data is allowed, including pre-trained models \n- Submission file must be generated by the API.\n\nSubmission runtimes have been obfuscated. If you repeat the exact same submission you will see up to 30 minutes of variance in the time before you receive your score.\n\nPlease see the [Code Competition FAQ](https://www.kaggle.com/docs/competitions#notebooks-only-FAQ) for more information on how to submit. And review the [code debugging doc](https://www.kaggle.com/code-competition-debugging) if you are encountering submission errors.", "isPublished": false, "hasIsPublished": false, "postTitle": "", "hasPostTitle": false, "mimeType": "", "hasMimeType": false }, { "name": "Upgraded Kaggle Hardware", "content": "We're excited to announce that we've added powerful new H100 machines to the AIMO hardware pool! See [this page](https://cloud.google.com/compute/docs/accelerator-optimized-machines#a3-high-vms) for more on these machines.\n\nWhat you need to know:\n\n- AIMO 3 Only - H100 are only available for notebooks attached to this competition, use of H100s for any other activity could result in moderation action (e.g. site suspension or account ban).\n- No Internet - To help ensure these machines are reserved for competition use, all H100 sessions must have internet disabled.\n", "isPublished": false, "hasIsPublished": false, "postTitle": "", "hasPostTitle": false, "mimeType": "", "hasMimeType": false }, { "name": "Note on Language and Notation", "content": "All problems are text-only with mathematical notation in LaTeX. The following conventions are observed:\n- \\\\(\\\\verb|amssymb|\\\\), \\\\(\\\\verb|amsmath|\\\\) and \\\\(\\\\verb|enumitem|\\\\) packages are used. No other package is assumed.\n- Variable names are given in math mode e.g. \\\\(a\\\\), \\\\(x\\\\), \\\\(y\\\\). \\\\(\\verb|\\mathcal|\\\\) letters may be used, e.g. \\\\(\\\\mathcal{R}\\\\), \\\\(\\\\mathcal{S}\\\\).\n- Integers such as 1, 2 can be outside of math mode or inside. Large integers such as \\\\(10^{50}\\\\) are in math mode.\n- \\\\(\\\\verb|itemize|\\\\) and \\\\(\\\\verb|enumerate|\\\\) environments may be used (with the latter using the conventions from the \\\\(\\\\verb|enumitem|\\\\) package).\n- \\\\(\\\\mathtt{\\\\$}...\\\\mathtt{\\\\$}\\\\), \\\\(\\\\verb|\\\\ \\(...\\\\ \\)|\\\\), \\\\(\\\\verb|\\\\\\[...\\\\\\]|\\\\), \\\\(\\\\verb|\\\\begin{equation}...\\\\end{equation}|\\\\), and \\\\(\\\\verb|\\\\begin{equation\\*}...\\\\end{equation\\*}|\\\\) may all be used for equations.\n- No environments other than the ones mentioned above are used.\n- \\\\(\\\\verb|\\\\ldots|\\\\), \\\\(\\\\verb|\\\\cdots|\\\\), \\\\(\\\\verb|...|\\\\) can be used to continue a finite sequence or infinite.\n- Multiplication can be represented by simple adjacency \\\\(2x\\\\) or using LaTeX \\\\(\\\\verb|\\\\cdot|\\\\) command for \\\\(x \\\\cdot y\\\\) or LaTeX \\\\(\\\\verb|\\\\times|\\\\) command for \\\\(x \\\\times y\\\\).\n- \\\\(n!\\\\) for an integer \\\\(n \\\\geq 0\\\\) represents the factorial function \\\\(n! = n\\\\cdot(n-1)\\\\cdot(n-2) \\\\cdots 2 \\\\cdot 1\\\\) (with \\\\(0!=1\\\\)).\n- \\\\(\\\\log\\\\), \\\\(\\\\sin\\\\), \\\\(\\\\cos\\\\) are generally in math mode. Unless a base is explictly stated (e.g., $\\log_{2}$), $\\log$ refers to the natural logarithm.\n- Fractions may be represented using \\\\(\\\\frac{a}{b}\\\\) or inline using \\\\(a/b\\\\).\n- \\\\(\\\\lvert x \\\\rvert\\\\) represents the absolute value of \\\\(x\\\\), e.g., \\\\(\\\\lvert -2 \\\\rvert = 2\\\\).\n- \\\\(x \\%\\\\), where \\\\(x\\\\) is a real number, represents the value \\\\(\\\\frac{x}{100}\\\\)\n- Standard geometric notation is used such as the triangle \\\\(\\\\triangle ABC\\\\) with sides \\\\(AB\\\\), \\\\(BC\\\\), and \\\\(AC\\\\). Standard notation is also used for angles, e.g., \\\\(\\\\angle ABC\\\\). Angles may be specified in degrees \\\\(\\\\angle ABC = 60^\\\\circ\\\\). Measures with no units should be assumed to be in radians.\n- Standard trigonometric operators including \\\\(\\\\sin\\\\), \\\\(\\\\cos\\\\), \\\\(\\\\tan\\\\) may be used.\n- Sets are denoted using standard curly bracket notation, e.g., \\\\(\\\\{1,2,3\\\\}\\\\).\n- \\\\(\\\\lfloor x \\\\rfloor\\\\) represents the integer part of the real number \\\\(x\\\\) ie the greatest integer less than or equal to \\\\(x\\\\). \\\\(\\\\lceil x \\\\rceil\\\\) represents the smallest integer greater than or equal to \\\\(x\\\\). \\\\(\\\\{x\\\\}\\\\) for a single value \\\\(x\\\\) refers to the fractional part of the real number \\\\(x\\\\) ie \\\\(x - \\\\lfloor x \\\\rfloor\\\\).\n- \\\\(\\\\verb|\\\\left|\\\\) and \\\\(\\\\verb|\\\\right|\\\\) may be used in conjunction with other LaTeX commands, e.g. \\\\(\\\\verb|\\\\left \\\\lceil \\\\frac{x}{2}\\\\right\\\\rceil|\\\\) producing \\\\(\\\\left \\\\lceil \\\\frac{x}{2}\\\\right\\\\rceil\\\\).\n- Overline notation \\\\(\\\\overline{abcd}\\\\) represents the integer formed by the digits \\\\(a\\\\),\\\\(b\\\\),\\\\(c\\\\),\\\\(d \\\\in \\\\{0, 1, \\\\ldots, 9\\\\}\\\\) with \\\\(a \\\\neq 0\\\\), e.g. with \\\\(a=1\\\\), \\\\(b=2\\\\): \\\\(\\\\overline{ab}=12\\\\) and \\\\(\\\\overline{bab}=212\\\\).\n- \\\\(\\\\binom{a}{b}\\\\) is interpreted as the binomial coefficient, representing the number of ways of choosing \\\\(b \\\\geq 0\\\\) elements from a set of size \\\\(a \\\\geq 0\\\\). We use the convention that \\\\(\\\\binom{0}{0}=1\\\\) and \\\\(\\binom{a}{b}=0\\\\) if \\\\(b>a\\\\).\n- \\\\(\\\\mathbb{N},\\\\mathbb{Z},\\\\mathbb{Q}\\\\), and \\\\(\\\\mathbb{R}\\\\) are used to denote the set of positive integers, integers, rational numbers, and real numbers, respectively (where positive means \\\\(>0\\\\)). For an integer \\\\(n\\\\), the notation \\\\(\\\\mathbb{Z}_{\\\\geq n}\\\\) may be used to denote the set of integers \\\\(\\\\geq n\\\\).\n- Standard notation for functions will be used such as \\\\(f\\colon S \\to T\\\\) where \\\\(S\\\\) and \\\\(T\\\\) are sets.\n- We allow informal probability language such as: a point is chosen uniformly at random in the interval \\\\([0, 1]\\\\).\n- The sum over the empty set is \\\\(0\\\\), and the product over the empty set is \\\\(1\\\\).\n- For integers \\\\(l\\\\), \\\\(m\\\\), and \\\\(n\\\\), \\\\(l^{m^n}\\\\) denotes \\\\(l^{\\\\left(m^n\\\\right)}\\\\).\n- British or American versions of English can be used. Thus ''highest common factor'' means the same as ''greatest common divisor''.\n- A prefix or subscript may be used to indicate features of a triangle associated with vertices. Thus triangle \\\\(ABC\\\\) has three altitudes, and the one dropped from \\\\(A\\\\) could be denoted the altitude through \\\\(A\\\\), the \\\\(A\\\\)-altitude or the altitude \\\\(h_a\\\\). Similarly, for median lines.\n- \\\\(:=\\\\) means 'is defined to be equal to'.\n- Taxonomy is *Bourbakist*: equilateral triangles are isosceles triangles, squares are rectangles and rectangles are parallelograms.\n- The word *trapezium* can mean different things in different countries. For the purposes of this competition, a trapezium is a quadrilateral with at least one pair of parallel opposite sides.\n\n**Please read this section carefully as there have been changes from the answer format used for AIMO2. We recommend looking through the reference problems to see examples.**\n\n- All answers must be reported as integers in the range \\\\(0\\\\) to \\\\(99999\\\\) inclusive. \n- All questions are worded so that the final answer will lie in this range (meaning any modulus required is explicitly given in the problem statement).\n- For questions of the form: `What is the remainder when \\\\(N\\\\) is divided by \\\\(q\\\\)?' where \\\\(q\\\\) is a positive integer and \\\\(N\\\\) is an integer, you should return the unique integer \\\\(r\\\\) such that \\\\(0 \\\\leq r