Smartest Animals: How We Measure Animal Intelligence

Animal intelligence is measured through observable behaviors including problem-solving, tool use, social learning, memory, and future planning, with great apes, dolphins, elephants, and corvids consistently performing at the upper range across multiple cognitive tests. No single metric ranks species on one intelligence scale because each animal evolved cognitive abilities suited to its ecological niche, and researchers use diverse methods to compare mental capacities across species that differ enormously in sensory systems, social structure, and physical environment.

This article explains the main approaches scientists use to measure animal intelligence, describes what those methods reveal about the cognitive abilities of several well-studied species, and provides a practical framework for evaluating intelligence claims. The content is written for students, researchers, life-science professionals, and informed general readers who want to understand how comparative cognition research works and what its limitations are.

What Animal Intelligence Means in Research

Intelligence in animals refers to the capacity to acquire, process, store, and use information to solve problems or adapt behavior to new circumstances. Researchers study this capacity through controlled experiments that measure specific cognitive functions instead of through a single intelligence quotient equivalent to the human IQ test.

The scientific study of animal cognition examines several distinct abilities. Problem-solving involves finding novel solutions to obstacles or puzzles. Tool use means manipulating an external object to achieve a goal that would otherwise be difficult or impossible. Social learning describes the ability to acquire new behaviors by observing others. Memory includes both short-term recall and long-term retention of information. Future planning refers to the capacity to anticipate future needs and act accordingly. Each of these abilities can be measured independently, and a species may excel in one domain while showing modest performance in another.

Comparative cognition researchers emphasize that intelligence is not a single trait distributed along one continuum from simple to complex. A honeybee that navigates using celestial cues and communicates the location of food sources through a symbolic dance demonstrates sophisticated cognitive abilities within its ecological context, even though it would fail most tests designed for mammals. The question of which animal is smartest therefore depends on which abilities are being measured and how those measurements are interpreted.

Core Principles of Measuring Animal Intelligence

Controlled Experiments and Standardized Tasks

Cognitive research with animals relies on controlled experiments that isolate specific mental processes. The T-maze is a classic apparatus used to assess memory and learning in rodents. The maze is shaped like a T placed horizontally, with the animal starting at the base and choosing one of the two goal arms at the other end. When two trials are given in quick succession, the rodent tends to choose the arm it did not visit before, a behavior called spontaneous alternation that reflects memory of the first choice. This tendency can be reinforced by making the animal hungry and rewarding it with preferred food when it alternates. Both spontaneous and rewarded alternation are sensitive to dysfunction of the hippocampus, though other brain structures are also involved. Each trial should be completed in under two minutes, but the total number of trials required varies according to statistical and scientific requirements. This protocol, detailed in Nature Protocols, demonstrates how a simple apparatus can yield reliable measurements of memory function.

Standardized tasks allow researchers to compare performance across individuals and species. The key requirement is that the task must be equally understandable and physically possible for all animals being compared. A task that requires fine manual dexterity will disadvantage species without manipulative appendages, while a task requiring visual discrimination will disadvantage species that rely primarily on olfaction.

Encephalization and Brain Structure

One approach to measuring intelligence involves examining brain size relative to body size. Encephalization implies an increase in brain or neocortex size relative to body size, size of lower brain areas, or evolutionary time. A review of 26 large-scale comparative studies in Progress in Brain Research provides robust evidence for five lifestyle correlates of encephalization: group living, a large home range, a high-quality diet, a strong reliance on vision, and arboreal or forest dwelling. The same review identifies six cognitive correlates: better performance in captive tests, more tactical deception, innovation, tool use, social learning, and general intelligence. Encephalization also correlates with a longer lifespan and with evolutionary changes including a high rate of change in microcephaly genes and an increase in brain size over macroevolutionary time.

The review notes that corrected neocortex size, expressed either as a ratio or a residual, is the most popular structural correlate of functional variables, while residual brain size is the measure associated with the greatest number of cognitive correlates. Controversies remain about whether to use corrected or absolute measures of neural structure size, whether brain parts evolved in a concerted or mosaic fashion, and whether cognitive processes are supported by specialized or domain-general brain structures.

Brain structure measurements provide indirect evidence about cognitive capacity. They cannot tell researchers what an animal actually does with its brain, only what neural resources are available. A large brain may support complex cognition, but it may also reflect other factors such as enhanced sensory processing or greater motor control.

Observational Studies of Natural Behavior

Some intelligence measures come from observing animals in their natural habitats. Researchers document instances of innovation, such as a bird discovering a new foraging technique, or social learning, such as a juvenile acquiring a behavior by watching an adult. These observations provide ecological validity because they show how animals use cognitive abilities in real-world contexts.

Observational studies have limitations. They depend on chance encounters and cannot be easily replicated. They also cannot distinguish between behaviors that arise from individual insight and behaviors that arise from trial-and-error learning or innate predispositions. For these reasons, observational evidence is most valuable when combined with controlled experiments that test specific hypotheses.

At a Glance: Intelligence Indicators Across Species

The following table summarizes the main intelligence indicators studied in several well-known species and the evidence available for each. The table is intended for orientation, not as a definitive ranking.

Research Methods for Studying Animal Cognition

Problem-Solving Tasks

Problem-solving tasks present animals with obstacles that require novel solutions. A typical setup might involve a puzzle box with a food reward that can only be accessed by completing a sequence of actions, such as pulling a lever, sliding a latch, or rotating a door. Researchers measure the time to solution, the number of attempts, and whether the animal improves with repeated exposure.

These tasks reveal whether an animal can understand cause-and-effect relationships and plan sequences of actions. Success requires more than physical ability, the animal must represent the goal and select actions that lead toward it. Comparative studies show that performance on problem-solving tasks correlates with encephalization measures across species, supporting the idea that these tasks tap into general cognitive capacity.

Tool Use Studies

Tool use provides some of the most striking evidence for animal intelligence because it requires understanding that an external object can serve as a means to achieve a goal. Researchers study both spontaneous tool use in the wild and tool use elicited in controlled experiments.

The New Caledonian crow is famous for manufacturing hooks from twigs and leaves to extract insects from crevices. This behavior appears to involve understanding of physical properties such as rigidity and shape. Great apes use sticks to fish for termites, stones to crack nuts, and leaves as sponges to collect water. Dolphins have been observed carrying marine sponges on their beaks to protect themselves while foraging on the seafloor, a behavior that appears to be socially transmitted from mothers to daughters.

Tool use studies must distinguish between flexible, insightful tool use and rigid, stereotyped behaviors that happen to involve objects. An animal that uses the same tool in the same way every time may be following an innate program instead of demonstrating understanding. Researchers therefore look for evidence of flexibility, such as modifying tools to suit different purposes or selecting appropriate tools for different tasks.

Social Learning Experiments

Social learning experiments test whether animals can acquire new behaviors by observing others. A common design involves a demonstrator animal that has learned a particular technique, such as opening a food container in a specific way, and a naive observer that watches the demonstrator before being given access to the same container. If the observer uses the same technique as the demonstrator, social learning is inferred.

Social learning is a foundation for culture, the transmission of behaviors across generations through teaching and imitation. Comparative studies show that social learning correlates with encephalization, suggesting that the cognitive demands of learning from others contributed to brain evolution in some lineages.

Memory Tests

Memory tests measure the ability to retain and retrieve information over time. The T-maze alternation protocol described earlier is one example. Other tests involve delayed matching-to-sample, where an animal must remember a stimulus after a delay and select it from an array of alternatives, or cache recovery tasks, where food-storing birds must remember the locations of hundreds of hidden food items.

Long-term memory can be tested by exposing animals to individuals or locations they have not encountered for months or years and measuring recognition responses. Dolphins recognize the signature whistles of former companions after decades of separation, and elephants appear to remember water sources and migration routes over very long periods.

Future Planning and Temporal Cognition

A growing body of research examines whether animals can represent time and plan for the future. The question of whether non-human animals possess genuine temporal cognition, the capacity to mentally represent time instead of merely respond to temporal cues, is addressed in a comparative analysis across apes, marine mammals, terrestrial mammals, birds, insects, and human infants. This work challenges the long-standing view that non-human animals operate in a permanent present and lack the ability to mentally represent time. The evidence supports the view that temporal cognition is widespread in nature.

Future planning is tested through experiments where animals must take action in the present to secure a future benefit. For example, some birds cache food in locations where they will need it later, and some primates select and carry tools to a future foraging site. These behaviors suggest the capacity to anticipate future needs, though alternative explanations based on learned associations must be carefully ruled out.

Species Profiles: What the Evidence Shows

Great Apes

Great apes, including chimpanzees, bonobos, gorillas, and orangutans, are the closest living relatives of humans and show the most human-like cognitive abilities. They use tools in the wild, learn behaviors from group members, and solve novel problems in laboratory settings.

Encephalization research places great apes among the most encephalized primates, with brain sizes that are large relative to body size. The comparative review of 26 studies found that encephalization correlates with better performance in captive tests, more tactical deception, innovation, tool use, and social learning. Great apes show all of these behaviors to a high degree.

One notable finding is that great apes appear to understand the mental states of others, an ability called theory of mind. They can predict what another individual knows or believes and adjust their behavior accordingly. This capacity supports complex social strategies such as deception and cooperation.

Bottlenose Dolphins

Bottlenose dolphins have large brains relative to body size and live in complex social groups with individual recognition, alliances, and cooperative hunting. They show strong evidence for vocal learning, including the ability to imitate sounds and to learn individually distinctive signature whistles.

Dolphins solve problems in laboratory settings, including tasks that require understanding of symbolic communication. Some studies have shown that dolphins can understand artificial languages consisting of acoustic symbols combined according to grammatical rules. They also show evidence of self-recognition in mirrors, a capacity shared with great apes and elephants.

The social intelligence hypothesis suggests that dolphins evolved large brains to manage the demands of life in fluid, fission-fusion societies where individuals must track many social relationships and coordinate cooperative activities.

Corvids

Corvids, the family of birds that includes crows, ravens, jays, and magpies, show cognitive abilities that rival those of great apes despite having very different brain structures. New Caledonian crows manufacture hooks from twigs and leaves, and other corvid species solve multi-step puzzles that require planning.

Food-storing corvids, such as Clark's nutcrackers and scrub jays, hide thousands of food items and remember their locations with remarkable accuracy. They also show evidence of episodic-like memory, the ability to recall what, where, and when a particular event occurred. Scrub jays can remember which food items they cached in which locations and how long ago they cached them, allowing them to retrieve perishable items before they spoil.

Corvid research demonstrates that intelligence does not require a mammalian brain. Bird brains are organized differently from mammal brains, yet they support cognitive abilities comparable to those of primates in many domains.

Elephants

Elephants have the largest absolute brain size of any land animal and show complex social behavior, including cooperation, empathy, and apparent mourning of dead companions. They have strong long-term memory for locations, individuals, and social relationships.

Elephants use tools in limited ways, such as using branches to swat flies or to scratch themselves. They also show evidence of social learning, including the transmission of vocalizations and foraging techniques within family groups.

The cognitive abilities of elephants are difficult to study in controlled laboratory settings because of their size and the ethical considerations of keeping them in captivity. Much of the evidence for elephant intelligence comes from observational studies in the wild and from semi-captive populations where researchers can conduct structured observations.

Octopuses

Octopuses are mollusks that have evolved complex nervous systems independently of vertebrates. They show remarkable problem-solving abilities, including opening jars, navigating mazes, and escaping from enclosures. Their eight arms contain a large proportion of their neurons, and research on embodied cognition shows that some computational work is offloaded to the peripheral nervous system.

Octopuses use objects as tools, such as carrying coconut shell halves to use as shelters. They also show evidence of individual personality differences and of learning through observation, though social learning is limited because octopuses are largely solitary.

The octopus demonstrates that complex cognition can evolve in lineages very different from vertebrates. Their distributed nervous system challenges assumptions about how intelligence is organized in the brain.

Rodents

Rats and mice are the most commonly studied animals in cognitive research because they are small, easy to house, and have well-characterized genetics. The T-maze alternation protocol is a standard method for assessing memory in rodents, and it is sensitive to hippocampal function.

Rodents show evidence of spatial memory, social learning, and some forms of planning. They can navigate complex mazes, remember the locations of food rewards, and learn from observing conspecifics. However, their cognitive abilities are generally considered more limited than those of primates, dolphins, and corvids in domains such as tool use and abstract reasoning.

Rodent research is valuable for understanding the neural mechanisms underlying cognition because researchers can manipulate genes, lesions, and pharmacological agents to test causal hypotheses about brain function.

Practical Assessment Steps for Evaluating Intelligence Claims

When evaluating claims about animal intelligence, whether in scientific literature or popular media, apply the following assessment steps.

Step 1: Identify the Specific Ability Being Claimed

Determine whether the claim refers to problem-solving, tool use, social learning, memory, future planning, or some other cognitive domain. A claim that an animal is smart is too vague to evaluate. A claim that an animal can remember the location of 500 cached food items is specific and testable.

Step 2: Examine the Experimental Design

Ask whether the study used controlled conditions, adequate sample sizes, and appropriate controls for alternative explanations. A study that shows an animal solving a puzzle must rule out the possibility that the animal succeeded through trial and error instead of insight. A study that claims social learning must rule out the possibility that the observer learned independently.

Step 3: Consider the Ecological Relevance

Ask whether the task is meaningful for the species being tested. A task that requires visual discrimination may be inappropriate for a species that relies primarily on olfaction. A task that requires manual manipulation may disadvantage species without manipulative appendages. The most convincing studies design tasks that are ecologically relevant while still being controlled.

Step 4: Compare Across Multiple Studies

Single studies can produce misleading results due to chance, methodological flaws, or idiosyncratic individual animals. Convergent evidence from multiple studies using different methods provides stronger support for a cognitive ability. The encephalization review in Progress in Brain Research synthesized 26 comparative studies to identify robust correlates, demonstrating the value of convergent evidence.

Step 5: Distinguish Evidence from Interpretation

Separate the raw observations from the interpretations placed on them. An observation that a crow bends a wire to make a hook is a fact. The interpretation that this behavior reflects understanding of physical causality is an inference that requires additional evidence. Researchers may disagree about how to interpret the same observations.

Records and Measurements in Cognitive Research

Cognitive research generates several types of records that are essential for evaluating claims.

Behavioral Data

Behavioral data include video recordings, automated tracking data, and observer ratings of animal behavior. These records document what the animal did, when it did it, and in what context. Behavioral data must be collected using standardized protocols to ensure reliability across observers and sessions.

Performance Metrics

Performance metrics quantify cognitive abilities. Common metrics include time to solution, number of errors, proportion of correct choices, and latency to respond. These metrics allow statistical comparison across individuals and conditions. The T-maze protocol specifies that each trial should be completed in under two minutes, providing a clear performance criterion.

Neural Data

Neural data include brain measurements such as size, structure, and activity patterns. Encephalization measures require careful standardization of brain and body weights, and different measures can produce different rankings. The review of 26 comparative studies found that corrected neocortex size and residual brain size are the most commonly used measures, but controversies remain about which measure is most appropriate.

Environmental and Social Records

Environmental and social records document the conditions under which animals live and were tested. Captive animals may perform differently from wild animals on cognitive tasks, and social history can affect performance. Researchers must record housing conditions, social group composition, and testing history to interpret results accurately.

Common Failure Patterns in Intelligence Research

Anthropomorphic Bias

Anthropomorphic bias occurs when researchers interpret animal behavior in human terms without sufficient evidence. Assuming that an animal that solves a puzzle understands the problem the way a human would may lead to overestimation of cognitive abilities. Careful experimental design must rule out simpler explanations based on associative learning or innate predispositions.

Clever Hans Effects

The Clever Hans effect, named after a horse that appeared to perform arithmetic but was actually responding to subtle cues from its handler, occurs when animals use unintended cues to solve tasks. Researchers must ensure that experimenters are blind to the expected outcome and that animals cannot access cues from humans or the environment.

Task Inappropriateness

Tasks that are inappropriate for a species can produce false negatives, making intelligent animals appear unintelligent. A species that evolved to navigate by smell may perform poorly on a visual discrimination task even though it has sophisticated cognitive abilities in other domains. Researchers must design tasks that are physically possible and ecologically meaningful for each species.

Small Sample Sizes

Cognitive research often uses small numbers of animals because of practical constraints. A single exceptional individual can skew results, and failure to replicate findings across individuals or laboratories can indicate that the original result was due to chance or idiosyncratic factors.

Publication Bias

Studies with positive results are more likely to be published than studies with negative results. This bias can create an inflated impression of animal cognitive abilities because failed replications and null results are underrepresented in the literature.

Limitations of Intelligence Measurement

No Universal Intelligence Scale

There is no single intelligence scale that can rank all animal species. Different species have different sensory systems, motor abilities, and ecological niches, making direct comparison difficult. A ranking that places great apes above corvids may reflect the types of tasks used instead of any true difference in cognitive capacity.

Domain Specificity

Cognitive abilities are often domain-specific. A species may excel at spatial memory while showing modest performance on social reasoning tasks. The concept of general intelligence, a single factor that predicts performance across many cognitive domains, is well-established in humans but its applicability to other species is debated.

Environmental Influences

Cognitive performance is influenced by environmental factors including nutrition, social experience, and stress. Animals raised in enriched environments typically perform better on cognitive tasks than animals raised in impoverished environments. These environmental effects complicate comparisons between populations and species.

Ethical Constraints

Research on animal cognition is subject to ethical constraints that limit the types of experiments that can be conducted. Studies that involve deprivation, stress, or invasive procedures require ethical approval and must balance scientific value against animal welfare. These constraints are appropriate but they limit the questions that can be addressed.

Welfare and Safety Context

Animal Welfare in Cognitive Research

Cognitive research must comply with animal welfare regulations that vary by jurisdiction. Researchers must obtain ethical approval before conducting studies, minimize distress, and provide appropriate housing and enrichment. The 3Rs principle, which stands for replacement, reduction, and refinement, guides the ethical use of animals in research. Artificial intelligence approaches are being developed to reduce animal use in toxicity testing and other applications, with in silico models showing accuracy comparable to animal tests for some endpoints.

Safety Considerations for Handlers

Researchers working with intelligent animals must consider safety. Large animals such as elephants and dolphins can injure handlers, and even smaller animals can bite or scratch. Standard operating procedures should include appropriate handling techniques, protective equipment, and emergency protocols.

Professional Escalation Criteria

Researchers should escalate concerns to supervisors or ethics committees when they observe signs of animal distress, when experimental procedures deviate from approved protocols, or when unexpected adverse events occur. Cognitive research should be halted if animals show signs of stress that cannot be managed through environmental modification.

Frequently Asked Questions

What is the smartest animal in the world?

There is no universally accepted answer because intelligence is measured through many different abilities and no single metric ranks all species. Great apes, dolphins, elephants, and corvids consistently perform well across multiple cognitive tests, but each species has unique strengths. A species that excels at social reasoning may perform poorly on spatial tasks, and vice versa. The question of which animal is smartest depends on which abilities are being measured.

How do scientists measure animal intelligence?

Scientists measure animal intelligence through controlled experiments that assess specific cognitive functions including problem-solving, tool use, social learning, memory, and future planning. They also study brain structure through encephalization measures and observe natural behavior in the wild. The T-maze alternation protocol for rodents is one example of a standardized cognitive test. Convergent evidence from multiple methods provides the strongest support for cognitive abilities.

Are dolphins smarter than chimpanzees?

Dolphins and chimpanzees both show sophisticated cognitive abilities, but they excel in different domains. Chimpanzees show strong evidence for tool use and social learning, while dolphins show strong evidence for vocal learning and social cognition. Direct comparison is difficult because the species have different sensory systems and motor abilities. Most researchers avoid ranking species on a single intelligence scale and instead describe the specific abilities each species possesses.

Can birds be as intelligent as mammals?

Yes, some birds show cognitive abilities comparable to those of mammals. Corvids, including crows, ravens, and jays, solve multi-step puzzles, use tools, and show evidence of episodic-like memory. These abilities are supported by brain structures that are organized differently from mammal brains but that appear to support similar cognitive functions. Bird intelligence demonstrates that complex cognition can evolve in lineages very different from mammals.

What is encephalization and why does it matter?

Encephalization refers to brain size relative to body size or to other measures such as the size of lower brain areas. Comparative studies show that encephalization correlates with cognitive abilities including innovation, tool use, and social learning. However, encephalization is an indirect measure of intelligence because it does not reveal how the brain is used. Some species with relatively small brains show sophisticated cognitive abilities, and some species with large brains show limited evidence for complex cognition.

Do animals plan for the future?

Evidence suggests that some animals can anticipate future needs and take action in the present to secure future benefits. Food-storing birds cache items in locations where they will need them later, and some primates select and carry tools to future foraging sites. Research on temporal cognition challenges the view that animals operate only in the present and supports the idea that some species can mentally represent time.

How does artificial intelligence research relate to animal intelligence?

Artificial intelligence research sometimes draws inspiration from animal cognition, and some researchers use animal-inspired tasks to evaluate AI systems. The Animal-AI Environment is a virtual laboratory designed for comparative cognition and artificial intelligence research. However, artificial intelligence and animal intelligence are different phenomena, and comparisons between them require careful definition of terms and methods.

Why is it difficult to compare intelligence across species?

Comparing intelligence across species is difficult because species have different sensory systems, motor abilities, and ecological niches. A task that is appropriate for one species may be impossible for another for reasons unrelated to intelligence. Additionally, cognitive abilities are often domain-specific, so a species may excel in one area while showing modest performance in another. These factors make it impossible to create a single intelligence ranking that is meaningful across all species.

Related Articles

• Animal Research
• Research Figures: Making Methods and Results Easier to Evaluate
• Circulation Research
• Circulation Research
• Circulation Research

References and Further Reading

• NCBI Literature Resources. National Center for Biotechnology Information.
• PubMed. National Library of Medicine.
• Recent Advancements in Pathogenesis, Diagnostics and Treatment of Alzheimer's Disease.. Current neuropharmacology, 2020.
• T-maze alternation in the rodent.. Nature protocols, 2006.
• A new era of antibody discovery: an in-depth review of AI-driven approaches.. Drug discovery today, 2024.
• Artificial Intelligence That Predicts Sensitizing Potential of Cosmetic Ingredients with Accuracy Comparable to Animal and In Vitro Tests-How Does the Infotechnomics Compare to Other "Omics" in the Cosmetics Safety Assessment?. International journal of molecular sciences, 2023.
• Introduction to Special Issue: Computational Toxicology.. Chemical research in toxicology, 2021.
• Use of artificial intelligence in animal experimentation: A review.. Toxicology letters, 2025.
• Primate encephalization.. Progress in brain research, 2012.
• Neural and molecular investigation into the paraventricular thalamus for chronic restraint stress induced depressive-like behaviors.. Journal of advanced research, 2025.
• NEWS 1
• Betting is Loving and Bettors are Predators: A Conceptual Metaphor Approach to Online Sports Betting Advertising.. 2018.
• 'Social, innovative and smart cities are happy and resilient': insights from the WHO EURO 2014 International Healthy Cities Conference.. 2015.
• Analysis of Households' E-Waste Awareness, Disposal Behavior, and Estimation of Potential Waste Mobile Phones towards an Effective E-Waste Management System in Dubai.. 2021.
• Smart dental materials for antimicrobial applications.. 2023.
• Causal Discovery Combining K2 with Brain Storm Optimization Algorithm.. 2018.
• Temporal Cognition in Animals. 2026.
• Cognition Beyond Representation: Varieties of Situated Cognition in Animals. 2018.
• Evidence for future cognition in animals. 2012.
• How do animals navigate the urban jungle? A review of cognition in urban-adapted animals. Current Opinion in Behavioral Sciences, 2022.
• Prospective cognition in animals.. Behavioural Processes, 2009.
• Animals, zombanimals, and the total turing test: The essence of artificial intelligence. Journal of Logic Language and Information, 2000.
• Study on quantitative analysis of animal hair blended fiber content based on artificial intelligence. Journal of Instrumental Analysis, 2026.
• Human intelligence and turing test. AI and Society, 1998.
• Intelligence Tests for AI. Lecture Notes in Computer Science, 2025.
• The Animal-AI Environment: A virtual laboratory for comparative cognition and artificial intelligence research. Behavior Research Methods, 2025.

This article is educational and does not replace institutional policy, professional advice, or applicable safety and regulatory requirements.