TruthfulQA
TruthfulQA is a benchmark to measure whether language models are truthful in generating answers to questions. It comprises 817 questions that span 38 categories, including health, law, finance and politics. The questions are crafted such that some humans would answer falsely due to a false belief or misconception, testing models' ability to avoid generating false answers learned from human texts.
Phi-3.5-MoE-instruct from Microsoft currently leads the TruthfulQA leaderboard with a score of 0.775 across 17 evaluated AI models.
Phi-3.5-MoE-instruct leads with 77.5%, followed by 
Granite 3.3 8B Instruct at 66.9% and Phi 4 Mini at 66.4%.
Progress Over Time
Interactive timeline showing model performance evolution on TruthfulQA
TruthfulQA Leaderboard
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 | Microsoft | 60B | — | — | ||
| 2 | 8B | — | — | |||
| 3 | Microsoft | 4B | — | — | ||
| 4 | Microsoft | 4B | — | — | ||
| 5 |  Nous Research | 70B | — | — | ||
| 6 | 70B | — | — | |||
| 7 |  Alibaba Cloud / Qwen Team | 15B | — | — | ||
| 8 |  AI21 Labs | 398B | — | — | ||
| 9 | 7B | — | — | |||
| 10 | Alibaba Cloud / Qwen Team | 33B | — | — | ||
| 11 |  Cohere | 104B | — | — | ||
| 12 | Alibaba Cloud / Qwen Team | 72B | — | — | ||
| 13 | Alibaba Cloud / Qwen Team | 32B | — | — | ||
| 14 | AI21 Labs | 52B | — | — | ||
| 15 | 8B | — | — | |||
| 16 | Alibaba Cloud / Qwen Team | 7B | — | — | ||
| 17 |  Mistral AI | 12B | — | — |
FAQ
Common questions about TruthfulQA.
More evaluations to explore
Related benchmarks in the same category
A challenging dataset of 448 multiple-choice questions written by domain experts in biology, physics, and chemistry. Questions are Google-proof and extremely difficult, with PhD experts reaching 65% accuracy.
A more robust and challenging multi-task language understanding benchmark that extends MMLU by expanding multiple-choice options from 4 to 10, eliminating trivial questions, and focusing on reasoning-intensive tasks. Features over 12,000 curated questions across 14 domains and causes a 16-33% accuracy drop compared to original MMLU.
All 30 problems from the 2025 American Invitational Mathematics Examination (AIME I and AIME II), testing olympiad-level mathematical reasoning with integer answers from 000-999. Used as an AI benchmark to evaluate large language models' ability to solve complex mathematical problems requiring multi-step logical deductions and structured symbolic reasoning.
Massive Multitask Language Understanding benchmark testing knowledge across 57 diverse subjects including STEM, humanities, social sciences, and professional domains
A verified subset of 500 software engineering problems from real GitHub issues, validated by human annotators for evaluating language models' ability to resolve real-world coding issues by generating patches for Python codebases.
Humanity's Last Exam (HLE) is a multi-modal academic benchmark with 2,500 questions across mathematics, humanities, and natural sciences, designed to test LLM capabilities at the frontier of human knowledge with unambiguous, verifiable solutions