Carnegie Learning combines cognitive science with AI to provide K-12 students with an and adaptive math and literacy tutoring platform, which has been adopted by thousands of schools across the United States.

Carnegie Learning: K-12 adaptive learning platform based on cognitive science

Core parameters and statistics of Carnegie Learning

Carnegie Learning's product system is built around the trinity of "cognitive model + adaptive engine + teacher dashboard". It is not a general AI question and answer system, but a disciplinary learning platform that is deeply bound to the American K-12 curriculum. Its adaptive engine is based on more than thirty years of cognitive psychology and AI tutoring system (Intelligent Tutoring System) research at Carnegie Mellon University. It is fundamentally different from purely data-driven recommended learning products in the underlying logic.

Parameter	Value
Product Positioning	K-12 Adaptive Mathematics and Literacy Learning Platform
Core products	MATHia (mathematics), Fast ForWord (reading and writing), LiveLab (teaching analysis), MATHstream (video streaming), ClearMath Elementary (elementary school mathematics)
Target Users	K-12 Schools, Districts, Teachers and Students
Core Technology	Cognitive adaptive learning engine, knowledge graph, student behavior modeling, natural language reasoning scoring
Supported languages	English, Spanish (including world language courses: French, German, Chinese, Italian, etc.)
Deployment method	Cloud SaaS (teacher + student), some support offline practice
Number of schools covered	Thousands of schools in the United States
Subject coverage	Mathematics (grades 6-12 and advanced mathematics), ELA (English Language Arts), literacy intervention, world languages
Latest version	Unpublished (continuous iteration)

Parameter Interpretation: Carnegie Learning's "adaptive" is not a simple upgrading or downgrading of question difficulty, but a modeling of students' problem-solving process based on a "cognitive model" - including which problem-solving strategy students choose, which step they stay for too long, and what type of conceptual misunderstandings are reflected in wrong answers. This enables the engine to not only push questions that are "more suitable for you", but also locate specific breakpoints in students' current cognitive structures and push corresponding micro-lessons for targeted repairs. This is the key difference in the underlying teaching logic between it and free learning platforms such as Khan Academy.

Carnegie Learning’s users and market recognition

Carnegie Learning’s market roots are in the U.S. K-12 public education system, and its adoption logic is fundamentally different from consumer-grade AI tools—purchasing decisions are made at the school district or school level and are driven by multiple factors such as curriculum standards, teaching effectiveness data, and federal education funding.

Scale of school adoption: Thousands of K-12 schools across the United States use Carnegie Learning’s products, covering urban public school districts, suburban schools, and charter schools (Charter Schools). MATHia is listed as a recommended or approved mathematics intervention tool by multiple state education departments, and has a high adoption rate especially in differentiated teaching scenarios in mixed-ability classes.

Academic Effectiveness Verification: Carnegie Learning’s products have undergone multiple rounds of third-party educational research evaluations. According to publicly available independent research reports, students who used MATHia improved their scores on standardized math tests more than a control group with traditional instruction, particularly in core subjects such as Algebra I and Geometry. Fast ForWord has been verified in multiple peer-reviewed journals to have a positive intervention effect on the literacy skills of students with dyslexia and language development delays, and has been included in the list of recommended tools for special education in many states.

Third Party Evaluation and Certification: MATHia has obtained Research-Based Design certification from organizations such as Digital Promise, indicating that its product design has a solid foundation in educational research. Carnegie Learning is also a participant in the ESSA (Every Student Succeeds Act) rating system, and its product effectiveness meets Tier 2 (Moderate Evidence) to Tier 3 (Promising Evidence) evidence standards, which allows school districts to use federal Title I and other funds for procurement.

Market Controversies and Points of Attention: The effectiveness of adaptive learning products is highly dependent on the quality of teacher implementation. Some education researchers pointed out that if teachers only use MATHia as a "screen time" replacement rather than deeply integrating it with daily teaching, the actual effect will be significantly lower than the data disclosed in the research report. In addition, the digital procurement wave in school districts across the United States during the epidemic brought a wave of growth. However, after the epidemic, some school districts reduced software subscriptions due to budget cuts. The actual renewal rate of Carnegie Learning has not been made public. It is recommended that purchasers directly request the retention rate data of currently active schools from the sales team.

Industry Benchmarking: In the K-12 AI adaptive learning track, Carnegie Learning is in direct competition with DreamBox Learning, i-Ready (Curriculum Associates), ALEKS (McGraw-Hill), and IXL Learning. Carnegie Learning’s differentiated advantage lies in its roots in cognitive science and MATHia’s in-depth modeling of students’ problem-solving processes, rather than simple presumptions of ability levels.

Benchmarking Dimensions	Carnegie Learning	DreamBox Learning	i-Ready	ALEKS
Core Subjects	Mathematics (6-12), Literacy Intervention	Mathematics (K-8), Reading (K-8)	Mathematics, Reading (K-8)	Mathematics (3-12), Chemistry
Cognitive model depth	Problem-solving strategy + path + time modeling	Knowledge status tracking	Standard-aligned adaptive testing	Knowledge space theory
Teacher Dashboard	LiveLab Real-time Intervention Suggestions	Teaching Insights Panel	Group Teaching Suggestions	Progress Reports
Tutoring	MATHstream Video MATHia Adventure	None	None	None
ESSA Levels of Evidence	Tier 2-3	Tier 2-3	Tier 2-3	Tier 2-3

Market Recognition Summary: Carnegie Learning enjoys a high reputation in the educational research community, and its product design is influential in cognitive science and AIED (AI in Education) academic circles. But in the actual K-12 procurement market, its market share is not the first in the industry - IXL and i-Ready have more advantages in large-scale school district procurement due to wider grade coverage and lower prices. Carnegie Learning's advantage scenario is the upper grade (grades 6-12) market that requires in-depth math tutoring and literacy intervention, rather than comprehensive coverage of K-5.

Carnegie Learning’s Cost Advantage

Carnegie Learning’s cost structure uses schools/school districts as the purchasing body and cannot be purchased directly by individual families. This is fundamentally different from most consumer-oriented AI education products. The following is a cost analysis based on public information and industry practices (subject to official real-time quotations).

C-side/Individual Home: Carnegie Learning does not offer individual subscription plans. Students must obtain access through their school or district. This means families can’t pay directly for access, like with Duolingo Plus or Khan Academy’s donation plans. For homeschoolers or families with additional tutoring needs, please contact the sales team directly to inquire about the availability of limited external access plans - as of this writing, no such plans have been officially disclosed.

B-side/School and District Procurement: Carnegie Learning uses an annual fee based on the number of students in its pricing model. According to estimates from the public education technology procurement database, the annual license fee per student is roughly in the following range (business confirmation is required):

Product Portfolio	Estimated Annual Fee/Student	What’s Included
MATHia Single Subject Mathematics (Grades 6-12)	$30-45	Platform license, teacher training, technical support
Fast ForWord Literacy Intervention	$35-$55	Adaptive Diagnosis + Practice + Progress Tracking
MATHia + Fast ForWord Combo	$50-$70	Dual-subject licensing, cross-product data integration
All-subject package (includes ELA, World Languages, and more)	$60-$90	Multi-subject access, priority support, custom reporting

Price influencing factors: The actual transaction price is affected by the following variables - the total number of students (the unit price of small school districts with less than 500 students is usually higher than that of large districts with 10,000 students), the length of the contract (discounts are available for multi-year contracts), whether additional professional services are required (teacher training workshops, curriculum mapping services, data migration support), and whether to purchase through an intermediary dealer (such as an education service intermediary).

Cost comparison with competing products: Based on the math software license fee per student per year, Carnegie Learning is in the upper range of the industry median. IXL costs around $15-25/student/year, i-Ready is around $20-35/student/year, and DreamBox is around $35-50/student/year. The price of Carnegie Learning corresponds to the complexity of its product in terms of functional depth (cognitive model + micro-course + LiveLab + video coaching), but if you only need basic adaptive exercises, competing products provide more economical options.

Hidden Cost Analysis:

• Implementation Cost: Teacher training is the largest hidden investment. The depth of MATHia's cognitive model means that teachers need to invest time in understanding its teaching logic and data panels, otherwise it will be difficult to be effective. It is recommended that 3-5 days of teacher professional development be reserved in the purchasing budget.
• Curriculum Alignment Cost: While MATHia content is aligned to Common Core and state standards, districts adopting NGSS (Next Generation Science Standards) or other non-standard curriculum frameworks may require additional curriculum mapping work, which is typically charged a program fee based on district size.
• Infrastructure Cost: The SaaS model does not require a local server, but schools will need to ensure that students' Chromebooks, iPads, or Windows devices meet minimum browser and network bandwidth requirements. Because Fast ForWord involves audio processing, it has certain requirements for the quality of headphones and the audio delay of the device.

Main functions of Carnegie Learning

Carnegie Learning's functional matrix is not a single "AI counselor", but a four-layer product system designed around "cognitive diagnosis - personalized push - teacher intervention - effect evaluation". Here are its core features and synergies between them:

1. MATHia Cognitive Adaptive Mathematics Tutoring

Core Value: It is not about giving questions, judging right or wrong, and pushing the next question, but modeling students' "cognitive state" in real time and accurately locating conceptual gaps.

• Multi-skill level knowledge map: Each mathematical knowledge point (such as "One-dimensional linear equation") is broken down into 5-15 fine-grained skills (such as "Merge similar items", "Move items", "Coefficient to 1"). The system tracks students' mastery of each skill level and automatically blocks progression when they have not mastered prerequisite skills, ensuring that there are no structural holes in the learning path.
• Problem-solving strategy tracking: MATHia can not only determine whether the answer is correct or incorrect, but also identify which problem-solving strategy the student has adopted (such as "trial number method", "algebra method", "drawing method"), and determine the applicability of this strategy in the current problem situation. If a student consistently solves problems using the algebraic method instead of the algebraic method, the system will identify this problem and push guided micro-lessons.
• Natural Language Reasoning Scoring: In MATHia Writing, which requires students to explain problem-solving ideas in words, AI scores reasoning quality based on semantics rather than keyword matching. This solves the fundamental flaw of traditional adaptive learning systems that "only look at the results and not the process".
• Micro-lesson instant push: When the system detects that students repeatedly make mistakes in a certain skill, it will automatically insert a 3-5 minute interactive micro-lesson (built in MATHia instead of jumping to an external video) to explain the concept and immediately continue the practice, forming a "diagnosis-teaching-verification" concept.

Synergistic effect: MATHia's cognitive model data is directly input into LiveLab. Teachers can grasp the "cognitive map" of the whole class in real time without having to analyze students' homework themselves - which students are stuck in which sub-step of which skill, at a glance. This is fundamentally different from the "report card" style reporting of traditional LMSs.

2. Fast ForWord Cognitive Literacy Training

Core Value: Neurocognitive training for students with dyslexia and learning difficulties, rather than traditional "read more, write more" exercises.

• Adaptive Auditory Training: Train the brain's auditory processing speed by precisely adjusting the duration and frequency of speech stimulation. This is especially critical for dyslexic students with phonological processing deficit.
• Language and reading cognitive exercises: A layered exercise system from phonemic awareness to syntactic understanding and discourse reasoning. Each exercise section is designed based on the principle of brain plasticity (Neuroplasticity) in cognitive neuroscience, rather than "questions on demand" based on grade standards.
• Progress Tracking and Intervention Report: The system automatically generates students' cognitive skill improvement curves, and teachers can compare standardized assessment scores before and after training.

Synergy: Fast ForWord and MATHia can be connected at the data level - when LiveLab finds that a student's difficulty in mathematics is due to insufficient reading comprehension (the inability to understand the text of word problems), it can recommend that the teacher include the student in the Fast ForWord training plan simultaneously. This cross-product diagnostic linkage is a unique advantage of Carnegie Learning’s product matrix.

3. LiveLab teacher real-time dashboard

Core Value: The transformation of teaching decision-making from "judging who is good based on experience" to "knowing who is stuck where by looking at the data".

• Heat map of the whole class: Displays each student's current learning activities, mastery progress, and distribution of difficult knowledge points in real time. Green means "progressing", yellow means "encountering difficulties", and red means "stuck for a long time".
• Hierarchical grouping suggestions: The system automatically recommends a grouping plan of 3-5 people suitable for group teaching, clustering based on the students' current difficulty points - not a simple "group of good students/group of poor students", but "a group of students who are stuck on fraction multiplication, and a group of students who are stuck on geometric proofs".
• Intervention Action Record: Teachers can mark implemented intervention measures (such as one-on-one tutoring, re-explanation, additional exercises) in LiveLab, and the system will follow up on the effect of the measure to form a data-driven teaching intervention system.

4. MATHstream Interactive Video Tutoring

Core Value: Embed problem-solving videos of real teachers into adaptive learning paths to solve the problem of "pure AI tutoring lacking human touch".

• On-demand video push: When students are stuck on a certain knowledge point in MATHia, the system pushes a 2-4 minute problem-solving video recorded by a real mathematics teacher. The style is similar to TikTok/Reels, adapting to the short video consumption habits of contemporary students.
• Embedded Checkpoint: The video will be paused during playback and a check question will pop up. Students can only continue watching after answering correctly, ensuring that passive viewing does not become "background sound".

5. ClearMath Elementary elementary school math solutions

Core Value: Extend MATHia's cognitive model to the K-5 math stage, covering the complete learning path from number sense, addition and subtraction to fractions and decimals.

6. Other subjects and language courses

• ELA Curriculum: Mirrors & Windows (English Language Arts Curriculum), Clearfluency (Oral Fluency Assessment), Bookshop Phonics (Phonics).
• World Language Courses: Complete course systems such as Spanish, French, German, Chinese, Italian, etc. Each course provides teaching materials, online exercises and assessment tools.

Carnegie Learning’s model and version evolution

Carnegie Learning's product evolution is not a software iteration marked by "version number", but a development history based on product line expansion and cognitive engine upgrades. Understanding this evolution helps purchasers evaluate the maturity of the platform and the stability of the technology roadmap.

The first stage: the birth and deep cultivation of MATHia (1998-2014)

• 1998: The project begins as the Cognitive Psychology Research Program at Carnegie Mellon University, with a founding team including Steven Ritter, William S. Hadley, John R. Anderson (creator of the ACT-R cognitive architecture), and Kenneth Koedinger (an expert in human-computer interaction and learning science).
• 2011: Apollo Group acquires Carnegie Learning for $75 million and pays an additional $21.5 million to license related technology from Carnegie Mellon University.
• 2014: MATHia (originally named Cognitive Tutor) has been deployed maturely in public schools, covering a complete curriculum system from sixth grade to advanced mathematics. The cognitive model has achieved high accuracy after more than ten years of classroom teaching data accumulation.

The second phase: M&A expansion and discipline expansion (2015-2020)

• 2015: Apollo Group sells Carnegie Learning to a Chicago-based education investment group, returning the company to an independent path focused on K-12.
• 2018: Private equity firm CIP Capital acquired Carnegie Learning. In the same year, it integrated the publishing assets of New Mountain Learning and acquired textbook brands such as EMC, Paradigm, and JIST to complete subject lines such as ELA and World Languages.
• 2020: Acquired Scientific Learning Corporation and incorporated Fast ForWord literacy training products into the product matrix, marking a key step in the transformation from "mathematics as a single subject" to "all subjects".

The third phase: AI upgrade and product matrix improvement (2020-2026)

• 2020-2022: The epidemic drives a wave of digital procurement in districts across the United States. Carnegie Learning accelerates MATHia’s AI engine upgrade—introducing a more fine-grained policy tracking model, a natural language reasoning scoring module, and a real-time grouping algorithm.
• 2022: Acquires MUSE Virtual, a K-12 online learning platform founded by Suzy Amis Cameron and James Cameron, to gain virtual school operations experience and technology assets.
• 2023-2024: Launch of MATHstream interactive video tutoring MATHia Adventure (gamified learning) and ClearMath Elementary (elementary school mathematics program). The LiveLab dashboard has completed a major UI upgrade, adding cross-product data integration capabilities.
• October 2024: Opening of Canadian headquarters in St. John's, Newfoundland, Canada, marking the launch of expansion into markets outside of North America.
• 2025-2026: Continue to upgrade the deep learning model of the adaptive engine to embed NLP capabilities more deeply into the reasoning scoring and policy identification modules. LiveLab introduces predictive intervention suggestions - not only telling teachers "who is having trouble now" but also predicting "who may be left behind next week if no intervention is made."

Release Notes: Product version numbers for Carnegie Learning are not available to the public. The above milestones are based on Wikipedia records, official press releases and public industry reports. The precise engine version number and API version need to be obtained through commercial channels. It is recommended that the purchaser request the product update roadmap (Roadmap) for the last three years from the sales team as a basis for evaluation.

Carnegie Learning’s Technical Advantages

Carnegie Learning's technical barrier lies not in "what latest AI model is used", but in its in-depth understanding of educational scenarios and continuous polishing of cognitive models. The following dismantles its technical mechanism from four dimensions:

1. Cognitive model-driven adaptive engine (not data-driven)

Most adaptive learning systems (including some AI education products) adopt the "Knowledge Tracing" method - through a Bayesian model or a deep learning network, the probability of mastery is calculated based on the student's answer sequence. Carnegie Learning's MATHia engine adds three layers of modeling dimensions on this basis:

• Strategy Selection Model: Not only depends on "whether you can", but also "how you think". The system has a built-in problem-solving strategy library marked by multiple mathematics education experts. When students solve problems, the engine infers the strategy they adopt by analyzing the student's input sequence (such as moving items first or merging similar items first), and determines the effectiveness of the strategy in the current problem.
• Time Dimension Modeling: The dwell time of each step is incorporated into the model. An unusually long dwell time may indicate "guessing" or "distraction", and an unusually short time may indicate "random filling in" or "plagiarism". These signals are used to adjust the push strategy—increasing prompt frequency for guessing students and switching question presentation methods for distracted students.
• Error classification model: The system does not simply record "wrong answer", but classifies errors into types such as "concept misunderstanding", "calculation error", "improper strategy" or "question reading deviation". Different types of errors trigger different teaching responses.

Mechanism → Effect: The superimposed effect of the above three-layer model is that MATHia's treatment of a student who "makes a mistake in adding fractions" is not to simply give 10 more fraction addition questions, but to determine whether "the student does not understand the concept of common denominators (concept misunderstanding)" or "can use methods but frequently misses the steps of adding molecules (calculation errors)." The former will push micro-lectures on concept explanations, and the latter will push targeted exercises.

2. In-depth binding of knowledge map and curriculum standards

MATHia's content model is not just a skill-level tree, but is fine-grainedly aligned with Common Core State Standards (CCSS) and state curriculum standards. Each knowledge point corresponds to a set of CCSS standard codes in the system, which means:

• Standard Audit Facilities: Schools can accurately track "how well we are doing on CCSS.7.EE.B.4 (solving real-life, multi-step problems)."
• Flexible Curriculum Mapping: If a school district uses non-CCSS standards (such as TEKS for Texas), Carnegie Learning provides a standard custom mapping tool, but this tool requires certain configuration investment.
• Cross-grade connection: The system can identify that "a certain eighth-grade student's difficulty in solving equations is because the transfer skills of seventh-grade students have not been mastered" and automatically fill in the seventh-grade level content - this is a kind of curriculum adaptation flexibility in schools that uniformly teach by grade.

3. LiveLab’s data analysis and intervention guidance

LiveLab is not a traditional "student performance dashboard", but a Teaching Decision Support System:

• Real-time grouping algorithm: Based on the current learning status data of all students, a clustering algorithm is used to group students according to difficult knowledge points. Unlike grouping by performance, this dynamic grouping may change from class to class because it reflects "now" difficulty points rather than "last exam" results.
• Intervention Effect Tracking: After the teacher marks an intervention action, the system automatically tracks the student's performance changes in subsequent learning. Data accumulated over time can help teachers identify which intervention strategies are most effective for their own classes.
• Predictive Analytics: The latest version of LiveLab introduces simple time series forecasting - predicting a student's likely performance in the next unit assessment based on the current learning trajectory, helping teachers intervene in advance instead of remediating after the fact.

4. Differences in technical positioning from pure AI learning products

Technical dimension	Carnegie Learning	General AI assistants (such as ChatGPT/Claude) applied to education
Authoritative teaching content	Written by subject experts and aligned with curriculum standards	Model-generated content, not guaranteed by curriculum standards
Student model	Multi-dimensional cognitive model (strategy + time + error classification)	Unstructured student model, no memory of learner status
Teaching strategy	Hierarchical teaching strategy based on cognitive science	One-time question and answer, no long-term learning path planning
Data management	Diagnosis→Teaching→Practice→Assessment→Re-diagnosis	Each conversation is independent, no teaching management
Privacy compliance	FERPA compliant, COPPA compliant, data is not used for model training	The data usage policy of the free version and the commercial version needs to be distinguished

Summary of technical advantages: Carnegie Learning’s technical route determines that it is not the “fastest AI education product”, but it is likely to be the “adaptive platform with the most explainable teaching logic”. For education researchers and purchasing decision-makers, this explainability means that a product’s effects are reproducible and improveable—you can tell why it works and how to make it more effective.

How to use Carnegie Learning

Carnegie Learning's deployment and usage path is fundamentally different from consumer-facing AI products - it is a SaaS platform that requires unified deployment at the school/district level, rather than an application that individuals can register on their own. The following is the usage process of different roles:

School/District Procurement Initiation Process

1. Needs Assessment: The district determines the subjects that need coverage (math, literacy ELA, world languages), grade ranges, and budget. It is recommended to apply for a Pilot trial first (usually 3-6 months, covering 2-3 classes), and then expand after the teaching effect is verified.
2. Contract Signing: Contact the Carnegie Learning sales team (official website contact form or phone) to get a customized quote. To be confirmed in the contract: license type (per-student vs. per-school), data storage location (in the US vs. supports Canada/other regions), FERPA compliance terms, teacher training delivery method (online/offline/hybrid), and technical support SLA.
3. Implementation and Training: Officially provides teacher training (Professional Development) services, usually including a 2-3 day startup workshop and ongoing online support. It is recommended to prepare an "Instructional Technology Coach" internally as the main interface with Carnegie Learning.
4. Technical Deployment: Carnegie Learning is a cloud SaaS model, and schools do not need to install servers. The IT department needs to ensure that: all student devices can access the *.carnegielearning.com domain name, the browser is the latest version of Chrome/Firefox/Safari/Edge, and Fast ForWord requires headphones and a stable Internet connection.
5. Course Mapping: The school curriculum coordinator (Curriculum Coordinator) collaborates with Carnegie Learning’s curriculum service team to map the school’s curriculum plan (Pacing Guide) to the standard content sequence in the system.

Daily usage process for teachers

1. Pre-class preparation: Teachers log in to LiveLab to view the heat map of the class that day to understand which students have difficulties in which knowledge points, and adjust the teaching plan accordingly - which students need one-on-one tutoring, which students can advance independently, and which students are suitable for group discussions.
2. Classroom Implementation:
   • Students log in to MATHia or Fast ForWord to complete personalized exercises; teachers can monitor progress in LiveLab in real time.
   • When LiveLab shows that the class has low mastery of a certain knowledge point, the teacher will pause self-study and switch to unified explanation mode.
   • For students who need additional help, teachers can instantly view the "students' complete solution path" in MATHia - not only see the final answer, but also the student's sequence of operations at each step to quickly locate cognitive breakpoints.
3. Post-class evaluation: Use the learning reports generated by the system to conduct formative assessment (Formative Assessment) to identify knowledge points that require adjustment of teaching strategies.

Student entrance

Role	Entrance	Main Operations	Required Equipment
Students (Mathematics)	MATHia Web Application	Personalized exercises, micro-class learning, writing explanations	Chromebook/iPad/Windows Notebook
Students (Reading and Writing)	Fast ForWord Client/Web	Auditory training, language practice, reading practice	Computer + Headphones
Teachers	LiveLab Dashboard	Real-time monitoring, group management, intervention recording	Any web browser
Teaching researcher/principal	Management report panel	School-wide/district data overview, standard achievement report	Any web browser

Differences in user experience from general AI educational tools

The experience of using Carnegie Learning is not "opening a dialog box and asking AI questions", but a structured learning management system. Students cannot choose "what to study today" - the system's adaptive engine automatically assigns learning tasks based on cognitive models. The advantages and disadvantages of this design concept are equally obvious: the advantage is to ensure the integrity and systematicness of the learning path, but the disadvantage is that students lack the right to make independent choices, which may create a sense of restraint for students with strong self-motivation.

Product Pricing for Carnegie Learning

Carnegie Learning’s pricing system is centered on B-end school procurement and does not have a public price list. The following analysis is based on public education technology procurement data and industry practices. Actual quotations must be obtained through business channels.

Pricing Model Overview

Procurement model	Billing method	Applicable objects	Typical contract cycle
Single School License	Per Student Year Fee	Single K-12 School	1 Year
District License	Per-pupil-year fee (tiered discounts)	Districts with 2+ schools	1-3 years
Statewide	Custom Contracts	State Education Departments	3-5 Years
Pilot trial	Annual fee per student (usually 6-12 months)	Prospective school	6 months

C-side/individual

Carnegie Learning does not sell directly to individual households. This is an important decision limitation—homeschool families or students who want access to MATHia outside of class cannot pay directly like a subscription to Grammarly or Duolingo. As of the time of writing this article, no official individual subscription plans have been disclosed. Families in need can try to contact the sales team to inquire about whether exception arrangements exist, but this is not a standard product path.

B-side/schools and districts

Estimated price range (needs business confirmation):

Products	Annual Fee per Student (Estimated)	Services Included
MATHia (6-12 Mathematics)	US$30-45	Adaptive learning engine LiveLab, micro-course library, basic technical support
Fast ForWord (literacy intervention)	$35-$55	Adaptive diagnostics and training, progress tracking, teacher training
MATHia + Fast ForWord Combo	$50-$70	Cross-product data integration, unified LiveLab dashboard
Full Subject Bundle	$60-$90	Math + Literacy + ELA + World Languages + Priority Support

Room for Discount Negotiation: Based on industry practice in U.S. K-12 edtech procurement, large contracts ($100,000+ annual fee) typically receive discounts of 10%-25%. Multi-year contracts (3+ years) get an additional 5%-15% off the total price. Schools purchasing through the E-rate grant program also receive additional federal subsidies.

The most important price issue: Before signing the contract, be sure to ask the sales team to clarify whether the following costs are included in the quotation - the number and form of teacher training (Professional Development), the maximum working hours of the course mapping service, the cost of data migration (if migrating from other platforms), and the unit price for flexible expansion beyond the number of licensed students.

Enterprise/School District Level

Large school districts (more than 10,000 students) may use a customized contract model that includes:

• Dedicated Customer Success Manager
• Customized reports and data export (supports API/SFTP batch export)
• In-depth customization of teacher training (stratified training by subject/grade)
• Priority technical support response (e.g. response within 4 hours)
• Optional localized storage of student data (additional infrastructure fee negotiated)

Application scenarios of Carnegie Learning

The application scenarios of Carnegie Learning focus on the structured improvement of mathematics and literacy skills in the K-12 public education system. The following four scenarios have been verified through large-scale deployment:

Scenario 1: Differentiated teaching of mathematics in mixed-ability classes

Pain Point: In a seventh-grade mathematics class, one-third of the students have mastered the current knowledge points, one-third are just in a state of preparation, and one-third still need to supplement prerequisite skills. The traditional teaching at a unified pace leads to "the good ones can't get enough to eat, and the bad ones can't keep up."

Carnegie Learning’s Solution:

• MATHia assigns each student a personalized learning path, without requiring the entire class to study the same content at the same time.
• Teachers can master the "cognitive map" of the whole class in real time through LiveLab, and divide students who need similar help into groups for targeted tutoring.
• MATHstream short videos provide real-person guidance for stuck students, reducing the burden of teachers repeating explanations.

Cost reduction and efficiency improvement deduction:

• For teachers: From "correcting piles of homework to find out who can't" to "opening the panel to find out who is stuck where", the preparation time for differentiated diplomas has been reduced from 2-3 hours per week to 15-20 minutes (estimated). However, the interpretation ability of the LiveLab panel requires training, which may increase teachers’ cognitive load in the initial stage.
• For students: Students with a mastery of more than 80% can advance independently and no longer waste classroom time waiting for the whole class; students with insufficient mastery receive targeted remediation instead of being forced to follow content they do not understand.

Scenario 2: Dyslexia and special education literacy intervention

Pain Point: The incidence of dyslexia (Dyslexia) among K-12 students is about 5%-15%, but most digital learning products do not provide targeted cognitive training-they assume that students already have basic phonological processing abilities.

Carnegie Learning’s Solution:

• Fast ForWord's auditory training starts from the neurocognitive level and trains the brain's auditory processing speed by adjusting the speed and frequency of speech signals.
• Multiple standardized assessments before, during and after training to provide quantitative data support for the IEP (Individualized Education Plan).

Cost reduction and efficiency improvement: In traditional special education, a special education teacher usually covers 8-15 students, and most of them rely on paper-and-pencil exercises without computer assistance. Fast ForWord's automated diagnosis and training allows special education teachers to transform from "topic designers" to "training effect supervisors", allowing them to devote more time to one-to-one or group intervention. The field of special education usually pays more attention to intervention effect (Effect Size) than efficiency. The effect size (Cohen's d) reported by Fast ForWord in multiple peer-reviewed papers is in the range of 0.3-0.5, which is a moderately effective intervention - an acceptable level for educational intervention.

Scenario 3: Improvement of teaching quality at the school district level

Pain Point: School district teaching and research staff need to determine "which knowledge points are poorly mastered across the school district" in order to uniformly adjust the allocation of teaching resources. Traditional approaches rely on biannual standardized test data—lagged, sparse, and coarse-grained.

Carnegie Learning’s Solution:

• The management report panel provides four-level data drill-down capabilities by school district, school, grade, and class.
• Each knowledge point corresponds to a CCSS standard code, allowing teaching researchers to accurately identify "students in our district are lagging behind the state average on CCSS.8.F.A.1 (Function Definition Comprehension)."
• Data supports filtering by academic year, semester, quarter and customized time interval to track the changing trend of intervention effects.

Human-machine collaboration boundary: District-level data analysis can be completed automatically by the system - identifying weak knowledge points, comparing differences between schools, and tracking changing trends. However, the following sections require manual decision-making: whether to make unified teaching adjustments for the entire school district on weak knowledge points, whether to purchase additional supplementary teaching materials for this knowledge point, and how to allocate teacher training resources. The system provides answers to "what is it", but "what to do" still requires the professional judgment of teaching and research staff.

Scenario 4: Remediation of learning losses after the epidemic

Pain Points: The COVID-19 epidemic has caused significant learning losses (Learning Loss) in mathematics and reading among K-12 students across the United States, especially low-income school districts and minority students who are most affected.

Carnegie Learning’s Solution:

• MATHia's diagnostic features quickly identify gaps between a student's current mastery level and grade-level standards.
• The adaptive engine automatically replenishes prerequisite skills - if a ninth-grade student has a weak algebra foundation in seventh grade, the system will automatically arrange remedial exercises instead of directly pushing ninth-grade content.
• Teachers can implement "High-Dosage Tutoring" based on LiveLab data - targeted tutoring three times a week for more than 30 minutes each time with a teacher-student ratio of 1:3 or 1:4. Educational research believes this is one of the most effective forms of intervention to make up for learning loss.

Who is Carnegie Learning suitable for?

The applicable group of Carnegie Learning is not "all people who want to learn mathematics", but participants in specific educational scenarios. The following are the four core roles and their adaptation boundaries:

Applicable role one: K-12 mathematics teacher (grades 6-12)

• Core Value: From "correcting assignments to judge academic performance" to "real-time data-driven teaching". The MATHia + LiveLab combination liberates teachers' attention from "who got it right" to "who is stuck where, why, and what to do."
• Prerequisites: Basic digital teaching adaptability and willingness to incorporate data-driven decision-making into daily teaching are required. Teachers who are resistant to "screen teaching" may need a longer transition period.
• Misfit Boundary: If schools do not provide adequate teacher training time (at least 3 days of workshops + ongoing online support), the effectiveness of MATHia implementation will decrease significantly. Teachers cannot expect to "install it and it will automatically work" - the product is a teaching aid rather than a teaching substitute.

Applicable roles 2: Special education teachers and literacy intervention specialists

• Core Value: Fast ForWord provides quantitative data support for IEP goals and learning progress monitoring, reducing reliance on manual assessment. Automated auditory training allows students to complete daily training without the direct involvement of teachers, freeing up special education teachers’ time for more complex one-to-one intervention.
• Prerequisite: Students are required to be able to use computers and headphones independently to complete training (usually grade 2-3 or above). Students in younger grades or those with severe cognitive disabilities may need support from a paraprofessional.
• Not Fitting Boundaries: Fast ForWord is not the only or one-size-fits-all solution for dyslexia. It mainly targets reading difficulties at the phonological processing level and has limited effect on learning difficulties caused by attention deficit, visual processing disorders or contextual factors.

Applicable roles three: school district curriculum coordinators and teaching researchers

• Core Value: The ability to compare data across schools and grades enables the allocation of teaching resources at the school district level to shift from "experience-driven" to "evidence-driven".
• Prerequisite: Basic statistical literacy is required to interpret data reports and avoid over-interpretation of differences in small samples.
• Unfit Boundary: Although Carnegie Learning's data analysis dimensions (knowledge point mastery, strategy distribution, and progress speed) are rich, it does not provide causal inferences - such as "Why is school B's algebra score worse than school A"? This requires teaching researchers to make comprehensive judgments based on external information such as classroom observations, teacher status, and extracurricular resources.

Applicable role four: K-12 students (grades 6-12)

• Core Value: Obtain a truly personalized learning path - not a simple adaptation of "jump if you get it right, quit if you get it wrong", but a deep adaptation based on cognitive models. The learning process is not an isolated examination of questions, but a complete learning process of "diagnosis-micro-lecture-practice-re-diagnosis".
• Prerequisite: Students need basic computer operation skills and a certain degree of self-control. MATHia's learning process is structured and not suitable for exploratory learning of "learn whatever you want".
• Does not fit boundaries:
  • K-5 Students: MATHia's core design is for grades 6-12. Although ClearMath Elementary covers elementary school mathematics, the classroom interaction and game-based learning needs of this grade level do not fully match the structured style of MATHia.
  • Extraordinary students: For students who are extremely talented in mathematics and need to make leaps in learning progress, MATHia's pre-skill blocking mechanism may become a bottleneck - the system will insist on mastering all pre-requisite skills, even if the student has proven his ability through other means. However, teachers can adjust the learning path settings in the background to adapt to the needs of extraordinary students.
  • Non-US Curriculum Students: If a student's math curriculum is not based on Common Core or US State Standards, the cost of curriculum alignment may be higher than the value of the product itself.

Summary and Outlook of Carnegie Learning

Carnegie Learning occupies a unique position in the K-12 AI adaptive learning track - it does not pursue conversational fluency like a general AI assistant, nor does it pursue free and open access like Khan Academy. Instead, it deeply cultivates the "cognitive science + school procurement" segment, and has spent more than 20 years accumulating a set of explainable, verifiable, and academically endorsed mathematics and literacy tutoring systems.

Current Core Advantages:

• Cognitive Model Depth: Multi-dimensional modeling of students' strategy choices, problem-solving paths, time consumption and error types. This is a differentiated capability that most competing products do not have. For educators who value instructional logic over pacing efficiency, this depth means greater instructional interpretability.
• Academic Evidence Base: ESSA Tier 2-3 evidence levels, independent third-party research verification, and decades of classroom deployment data significantly reduce "compliance risk" when school districts purchase - evidence-based purchases with federal funds such as Title I, IDEA, and more.
• Product Matrix Completeness: From MATHia (math) to Fast ForWord (literacy) to LiveLab (data analysis) to MATHstream (video tutoring), Carnegie Learning has built a cross-disciplinary and cross-role product matrix rather than a single point tool. The linked diagnosis of mathematics and reading and writing data in LiveLab is a capability that is difficult to copy by competing products.

Major Current Limitations:

• Curriculum System Locked: The product is deeply bound to the U.S. Common Core and state standards. For schools using the IB (International Baccalaureate), Cambridge Curriculum, Chinese National Curriculum Standards, or any non-US curriculum system, curriculum alignment would be a huge investment and almost unfeasible. This limitation determines that Carnegie Learning's market penetration outside North America will be very slow.
• High purchase threshold: It is not sold to individual families, and there is no trial channel for self-service registration. School procurement decision cycles typically range from 6-18 months, which is not friendly to institutions that need to deploy quickly. The pilot trial process and pricing require contacting the sales team, and the initial process is not transparent enough.
• High teacher dependence: The effect of the product is highly dependent on the teacher's implementation quality - LiveLab's data requires teachers to actively interpret and use it, and MATHia's learning path requires teachers' understanding and cooperation. In schools with insufficient teacher training or frequent teacher turnover, the product may become an "expensive online operating system."
• Limited multi-language support: Although multiple world language courses are provided, the language and underlying content structure of the teaching interface are mainly English, which may create additional language barriers for students and families with limited English proficiency.

Follow-up observation points:

• Does the establishment of the Canadian headquarters mean the start of market expansion outside North America? Will there be content versions of the product adapted to Canadian, British or Singaporean curricula?
• Will the Predictive Analytics function introduce more machine learning models in future iterations, evolving from "describing learning conditions" to "recommending teaching strategies"?
• Will Carnegie Learning launch an individual/family subscription plan to cover the homeschool and after-school tutoring markets? This is one of the key gaps between it and the competition, such as IXL, which offers family plans.

Procurement and Adoption Risk Assessment: Carnegie Learning is not an educational product that "automatically increases points upon purchase" - its effectiveness is deeply tied to the quality of teacher training, school implementation capabilities and curriculum adaptability. For school districts considering procurement, it is recommended that the following evaluation paths be followed:

1. Pilot first: Select 2-3 representative classes (it is recommended to include a regular level class, a class with a high proportion of special education students, and a class with a high proportion of English learners) and run the Pilot for a full semester.
2. Controlled Experimental Design: The Pilot class uses MATHia, and the comparison class uses the existing teaching plan. At the end of the semester, the standardized test scores and classroom participation data of the two groups of students are compared.
3. Teacher Feedback Collection: Conduct structured interviews with participating teachers after the Pilot, focusing on understanding: Has LiveLab really changed teaching decisions? Is teacher training adequate? What are the barriers to daily use?
4. Full Cost Accounting: Before signing the procurement contract, calculate TCO (Total Cost of Ownership) = License Fee + Teacher Training Fee + Course Mapping Fee + IT Support Cost + Estimated Additional Teacher Time Cost.
5. Verification of key terms of the contract:
   • Student data storage location and FERPA/COPPA compliance terms (additional confirmation is required if Canadian data centers are used)
   • How the teacher training will be delivered, how many sessions it will take, whether it will include ongoing support or just a one-off workshop
   • Data export rights and export format upon termination of contract (whether to support common LMS standards such as 1EdTech)
   • Whether additional fees are required for major product updates, as well as the release frequency and notification lead time of updates
   • SLA terms - system availability, technical support response time, fault repair time (such as 99.5% uptime, 4-hour response on working days, etc.)
   • User expansion terms - if Pilot is successfully expanded to the entire school district, whether the unit price for new users will be locked or based on the current price list

One Sentence Decision: If your school district serves students mainly in grades 6-12, has serious math achievement gaps, has a team of teachers who are willing to embrace data-driven instruction, and has a 6+ month evaluation and deployment cycle - Carnegie Learning deserves to be on the Pilot short list. If you urgently need a solution covering K-5, or non-US curriculum systems, or have extremely high requirements for transparency in the procurement process (such as online self-service price comparison and instant activation), Carnegie Learning may not be the best choice.

Related tools: khanmigo, quizlet

Version Info

• Carnegie Learning latest version ：There is no official precise date yet, and the adaptive learning engine and course content will be continuously updated. （~2026）
• MATHia first edition ：There is no official precise date yet, Carnegie Learning launches AI math tutoring platform MATHia. （~2014）