TrustLLM is a framework and open-source toolkit for evaluating the trustworthiness of large language models. It was presented at ICML 2024 by an international team led by researchers in the United States. It proposes principles for trustworthy models and a benchmark covering truthfulness, safety, fairness, robustness, privacy and machine ethics. The benchmark draws on more than 30 datasets. A public leaderboard ranks 16 leading models. The code is released under the MIT licence.

AiEnsured is a commercial testing platform for AI products, designed to support the deployment of responsible AI. It is developed in India. The platform brings together fairness and bias evaluation, explainability, adversarial robustness testing and privacy and GDPR compliance checks. It also generates test cases automatically, including rare and extreme inputs. Further features cover experiment management, model comparison and performance testing. After deployment, it monitors models for concept drift, which can reduce accuracy over time.

The VeritasChain Protocol (VCP) is an open specification for cryptographically verifiable audit trails in algorithmic and AI-driven trading. It is maintained by the VeritasChain Standards Organization. VCP lets regulators and auditors verify mathematically that records of trading decisions, orders, risk controls and AI model use are complete and unaltered. A privacy module allows personal data to be erased while keeping the audit trail intact. VCP is designed to support MiFID II, EU AI Act and GDPR compliance.

KSAFE-MM is a Korean-language benchmark for evaluating safety risks in multimodal large language models. It was developed by K-intelligence at KT in South Korea. The benchmark contains 14,135 query-image pairs across 11 risk categories, including hate, violence, privacy and weaponisation. One subset tests Korean culture-specific risks and applies jailbreak strategies such as role-play. Researchers use it to check whether models respond safely. It is available for research under a CC BY-NC 4.0 licence.

IndoBias is a benchmark for evaluating social bias in large language models in Indonesian and three local languages. It was developed by researchers at the Mohamed bin Zayed University of Artificial Intelligence and Universitas Indonesia. One track tests whether models prefer stereotypical sentences over neutral ones. A second track tests whether models describe local groups and institutions more positively or negatively. The results show strong stereotypical bias in current models, especially regarding ideology and religion in local languages.

F2Bench is a benchmark for evaluating fairness in large language models that also considers factual accuracy. It was developed by researchers at Inner Mongolia University and Xiamen University. The benchmark contains 2 568 test instances across ten demographic categories, including intersectional combinations. One task tests whether models accept biased conclusions after gradual stereotypical suggestions in a conversation. A second task tests whether models avoid stereotypes while still reflecting real statistics accurately. The code and dataset are publicly available.

M4 Bias Eval FairFace is a dataset for evaluating social bias in vision-language models. It was created by Hugging Face to assess its IDEFICS models. It contains around 11,000 face images labelled by perceived gender, ethnicity and age. For each image, two models wrote a résumé, a dating profile and a news article about an arrest. Comparing these texts across demographic groups reveals stereotypes the models have learned. The dataset is released under a CC BY 4.0 licence.

FLEX is a benchmark for testing whether large language models stay fair when prompts are designed to induce bias. It was developed by researchers at Korea University. The benchmark contains 3,145 multiple-choice questions from established fairness datasets, each combined with an adversarial prompt. These include assigning negative personas, forbidding refusals and introducing small text changes. Results show that models appearing fair on standard benchmarks can still be easily manipulated. The data and code are available for research purposes.

CEB (Compositional Evaluation Benchmark for Fairness) is a fairness evaluation benchmark containing about 11 000 samples spanning multiple bias types, social groups, and tasks. It enables consistent measurement of compositional fairness behavior to identify failures that occur when fairness conditions combine across groups and bias factors. It is used by developers and auditors to support the fairness and accountability objectives, and to assess robustness of fairness performance under varied evaluation settings.

ASR-FairBench is an online leaderboard that ranks speech recognition systems on both accuracy and fairness. It was developed by researchers at IIT Kharagpur. The leaderboard tests models on recordings from speakers of different ages, genders, ethnicities and backgrounds. It measures how error rates differ across these groups and combines this with overall accuracy into a single score. Models are rated from "severely biased" to "exemplarily fair". Users can submit their own models and receive a fairness audit within minutes.

Audit the Judge is an open-source tool for checking whether large language models used as judges evaluate answers reliably. It builds answer pairs where the correct verdict is known. These test whether the judge invents preferences between equal answers or favours answers shown first or written at greater length. Results are presented in a one-page report with statistical confidence intervals. The tool works with any major LLM judge. It is written in Python and released under the MIT licence.

FairnessEval is an open-source framework for evaluating the fairness of machine learning models. It was developed by the University of Modena and Reggio Emilia with Microsoft. Users can load real or synthetic datasets and run fairness-aware models from toolkits such as Fairlearn and AIF360. The framework compares models on fairness, accuracy and training time, and presents the trade-offs in charts. It supports both selecting a suitable model and validating its performance under changing conditions.

Google's Differential Privacy libraries are an open-source collection of tools for producing statistics from sensitive data while protecting individuals' privacy. They add calibrated random noise to results such as counts and averages, so no single person can be identified. Building block libraries are available in C++, Go and Java. End-to-end frameworks make the tools usable by non-experts on large-scale data systems. Further tools track privacy budgets and audit privacy guarantees. The libraries are released under the Apache 2.0 licence.

FairX is an open-source toolkit for benchmarking machine learning models on fairness, data utility and explainability. It was developed by researchers at Linköping University. Users load tabular or image datasets and apply bias-mitigation methods before, during or after model training, including fair generative models that create synthetic data. FairX evaluates results using fairness, performance and synthetic data quality metrics. This lets users compare trade-offs between fairness and utility. It is written in Python and released under the MIT licence.

IEEE 3198-2025 is a technical standard that specifies a method for evaluating the fairness of machine learning systems. It was published by the IEEE Standards Association in May 2025. The standard categorises the causes of unfairness in machine learning and presents widely used definitions of fairness. For each definition, it specifies metrics and explains how to calculate them. Test cases set out detailed procedures for fairness evaluations. This supports consistent, repeatable fairness testing across systems and organisations.

Fairmetrics is an open-source R package for evaluating the group fairness of machine learning models. It was developed by researchers at the University of Toronto. The package calculates a wide range of fairness metrics, such as statistical parity and equal opportunity. Unlike most fairness tools, it also provides confidence intervals for each metric. This shows whether differences between groups are statistically significant or due to random variation. It is available on CRAN under the MIT licence.

MANILA is a web-based, low-code application for benchmarking machine learning models and methods that reduce bias. It helps users find the combination that offers the best balance between fairness and performance. MANILA models a fairness benchmarking workflow as a set of options, such as datasets, models and metrics. Rules between these options guide users and prevent experiments that would fail. This makes fairness benchmarking accessible to users with little or no programming experience.

DP-CGANs is an open-source Python library for generating synthetic data that protects individual privacy, designed with personal health data in mind. It was developed at Maastricht University. The library uses a conditional generative adversarial network, adapted to capture relationships between variables in the data. Differential privacy can be enabled during training, so no single person strongly influences the result. It works with tabular and RDF data. It is still under development and released under the MIT licence.

RecourseBench is a modular evaluation framework for algorithmic recourse methods that emphasises reproducibility when assessing user-facing counterfactual explanations. It enables practitioners to systematically compare recourse methods based on how they support actionable changes in decision-making systems. It is intended for researchers and developers who validate and improve the trustworthiness of explanation and human-agency features in AI used for consequential decisions.

Toolkit for evaluating fairness and bias in machine learning models using multiple subgroup fairness metrics (including parity and equalized-odds-style measures). It supports fairness auditing by quantifying disparities across demographic or other defined subgroups. Data scientists and developers can use it to verify and validate fairness properties and to guide improvements toward fairer model behavior.