ML model validation providers
31 published entries state ML model validation on their AI testing pages. Between them they name 8 AI product types.
Measurement of accuracy, calibration and error patterns of ML models, including metamorphic testing.
metamorphic testing: Checking that related inputs produce outputs related in a predictable way.Showing 31 of 31. Default order: coverage score, then name. How the score works.
| Rank | Company type | AI products (as stated) | Services | Verified | |||||
|---|---|---|---|---|---|---|---|---|---|
| 1 | DeviQA | Pure-play QA company | Poland, Warsaw | LLM applicationsAI agentsRecommender and ranking systemsPredictive ML modelsComputer visionVoice AI | 8 | 2010 | 16 | 2026-11-09 | |
| 2 | Abstracta | Pure-play QA company | United States, Jacksonville | LLM applicationsAI agentsPredictive ML models | 10 | 2008 | 15 | 2026-11-09 | |
| 2 | ImpactQA | Pure-play QA company | United States | LLM applicationsAI agentsComputer visionVoice AIRecommender and ranking systems | 9 | 15 | 2026-11-09 | ||
| 4 | BetterQA | Pure-play QA company | Romania, Cluj-Napoca | LLM applicationsAI agentsVoice AI | 9 | 2018 | 14 | 2026-11-09 | |
| 4 | TestDevLab | Pure-play QA company | Latvia, Riga | LLM applicationsAI agentsVoice AIComputer visionPredictive ML modelsRecommender and ranking systemsAutonomous and embedded AI | 5 | 2011 | 14 | 2026-11-09 | |
| 4 | TestingXperts | Pure-play QA company | United States | LLM applicationsAI agentsVoice AI | 10 | 14 | 2026-11-09 | ||
| 4 | Testlio | Crowdtesting platform | United States, Austin | LLM applicationsAI agentsPredictive ML modelsRecommender and ranking systems | 8 | 2012 | 14 | 2026-11-09 | |
| 4 | Testriq | Pure-play QA company | India, Mira-Bhayandar | LLM applicationsAI agentsPredictive ML models | 9 | 2010 | 14 | 2026-11-09 | |
| 10 | ITRex Group | IT services firm with a QA practice | United States, Sacramento | LLM applicationsAI agentsComputer visionPredictive ML modelsRecommender and ranking systems | 6 | 2009 | 13 | 2026-11-09 | |
| 11 | Fulminous Software | IT services firm with a QA practice | India, Jaipur | Voice AIAI agentsAutonomous and embedded AI | 7 | 2019 | 12 | 2026-11-09 | |
| 11 | QualityAI (formerly Qualitest) | Pure-play QA company | United States, Santa Clara | LLM applicationsAI agentsPredictive ML modelsRecommender and ranking systems | 7 | 12 | 2026-11-09 | ||
| 11 | SunTec India | Data and evaluation vendor | India, New Delhi | LLM applicationsDocument AIPredictive ML models | 7 | 1999 | 12 | 2026-11-09 | |
| 14 | Lionbridge | Data and evaluation vendor | United States, Waltham | LLM applicationsVoice AIComputer visionRecommender and ranking systems | 5 | 1996 | 11 | 2026-11-09 | |
| 14 | Testrig Technologies | Pure-play QA company | United States, Dallas | Predictive ML models | 8 | 2015 | 11 | 2026-11-09 | |
| 18 | Kualitatem | Pure-play QA company | United States, New York | Predictive ML models | 7 | 2010 | 10 | 2026-11-09 | |
| 18 | Sparkout Tech | IT services firm with a QA practice | India, Coimbatore | Predictive ML models | 7 | 2017 | 10 | 2026-11-09 | |
| 23 | BugRaptors | Pure-play QA company | India | Computer visionPredictive ML modelsVoice AI | 4 | 2016 | 9 | 2026-11-09 | |
| 23 | Chetu | IT services firm with a QA practice | United States, Sunrise | Predictive ML models | 6 | 2000 | 9 | 2026-11-09 | |
| 23 | Luxe Quality | Pure-play QA company | United States, McLean | Computer visionVoice AIPredictive ML models | 5 | 9 | 2026-11-09 | ||
| 29 | Frugal Testing | Pure-play QA company | India, Hyderabad | Voice AIComputer visionAutonomous and embedded AI | 4 | 8 | 2026-11-09 | ||
| 29 | SG Analytics | IT services firm with a QA practice | India, Pune | LLM applications | 5 | 2007 | 8 | 2026-11-09 | |
| 29 | TestMatick | Pure-play QA company | United States, New York | Recommender and ranking systemsPredictive ML models | 4 | 2009 | 8 | 2026-11-09 | |
| 34 | Cognizant | IT services firm with a QA practice | United States, Teaneck | AI agentsVoice AIPredictive ML models | 4 | 7 | 2026-11-09 | ||
| 34 | Macgence | Data and evaluation vendor | India, Noida | Predictive ML models | 5 | 7 | 2026-11-09 | ||
| 34 | MosChip | IT services firm with a QA practice | India, Hyderabad | AI agentsAutonomous and embedded AI | 3 | 1999 | 7 | 2026-11-09 | |
| 34 | Qualitrix | Pure-play QA company | India, Bengaluru | Predictive ML models | 5 | 7 | 2026-11-09 | ||
| 34 | Resillion | Pure-play QA company | United Kingdom | Predictive ML models | 5 | 7 | 2026-11-09 | ||
| 40 | QA Mentor | Pure-play QA company | United States, New York | Computer visionVoice AI | 2 | 2010 | 6 | 2026-11-09 | |
| 50 | QAlified | Pure-play QA company | Uruguay, Montevideo | LLM applications | 3 | 5 | 2026-11-09 | ||
| 50 | TestIT | Pure-play QA company | Hungary, Budapest | Predictive ML models | 3 | 5 | 2026-11-09 | ||
| 61 | Trilateral Research | AI assurance and audit specialist | United Kingdom, London | Predictive ML models | 3 | 4 | 2026-11-09 |
Product types covered by vendors stating this service
Count of entries on this page that state each product type.
| Product type | ML model validation |
|---|---|
| LLM applications | 14 |
| AI agents | 13 |
| Voice AI | 12 |
| Computer vision | 9 |
| Predictive ML models | 21 |
| Recommender and ranking systems | 8 |
| Generative media | 0 |
| Document AI | 1 |
| Autonomous and embedded AI | 4 |
Scope and standards referenced
Accuracy, precision, recall, F1, calibration, confusion analysis, metamorphic tests.
The directory lists which vendors state this service. It does not test whether a vendor applies these standards.
Related
Product types
Other services
Guides
Questions
What is machine learning model validation?
Model validation checks that a trained model performs as intended on data it has not seen: accuracy, precision, recall, calibration and behaviour on edge cases. 31 published entries state ML model validation.
Why is model validation important?
A model that scores well in training can fail on new data, on rare cases or after the data changes. Validation finds those gaps before release and gives the evidence that risk teams and regulators ask for.
How does the model validation process work?
Validators review the data and the intended use, test on held-out and edge-case data, compare metrics with thresholds, and probe robustness with perturbed or metamorphic inputs. The result is a report with findings and conditions for use.
When does it make sense to use an external model validation service?
When internal teams built the model and an independent check is wanted, when regulation or a customer asks for third-party evidence, or when the team lacks capacity for structured test design.
How does a model validation service differ from automated validation tools?
Tools compute metrics and run checks the team configures. A service designs the tests, chooses the data and judges the results for a specific use case. This directory lists services; tool-only vendors are on the not-published page.
How do you test a computer vision model?
Test on labelled images that match real conditions: lighting, angles, occlusion, image quality and rare classes. Measure accuracy per class and per condition, and check behaviour on altered images. 13 published entries state that they test computer vision.