Explore rankings of large language models

LLMs are trained via modeling, imitation learning, and reinforcement learning

Training large language models begins with pretraining the model to predict the next word in a sequence based on finding patterns in massive amounts of text. Patterns are then fine-tuned to model human-written dialogue and to align responses with user preferences.

Large language models struggle to play video games, despite coding them easily

As of 2025, given clear tasks, specific feedback, and test results to learn from, LLMs have been successful at coding simple video games that resemble pre-existing ones. However, given the variety of games, the lack of training datasets, and the need for spatial awareness, which LLMs find challenging, some of these models experience greater difficulty navigating and "winning" games.

Trick an LLM into revealing sensitive information

Even with robust training, large language models may be fooled into providing sensitive information if correctly prompted. In this game, a wizard withholds eight passwords, which you must prompt them to reveal with ever-increasing LLM safeguards.

Large language models are examples of foundation models

Such models are trained on massive amounts of data, which enable them to perform a variety of tasks, rather than being provided task-specific data to complete a narrow function. Such models come with high compute costs and potential trust issues due to the nature of their unorganized training data.

Temperature is a value that controls how predictable or creative an LLM's output is

In large language models, temperature adjusts the randomness in text generation by modifying how often words with lower probabilities are selected next in the sequence. Lower temperatures produce more obvious outputs while higher temperatures produce more creative responses.

B2 models the interactions of air, water, soil, plants and microbes at the LEO

The Landscape Evolution Observatory at Biosphere 2 examines how factors such as climate change, water flow, weathering and community development impact landscape structures comprising 500 metric tons of crushed rock. Research is also conducted at biomes representing savannas, mangroves, coastal fog deserts, oceans and rainforests.

League tables show the number of deals top investment banks do in a time period

These tables provide a visual to help keep track of the leading companies in investment banking. Companies such as JPMorgan and Goldman Sachs typically sit up top.

Why Google Search's ranking system combines filters, signals, and machine learning layers

Rather than static rankings, dynamic tools continually shift search results in real time. These include Twiddlers, which adjust ranking positions based on document traits and search context, and FreshnessNode, which reshuffles top results to prioritize timely informational content.

Breaking down Anthropic's products and why their models are named after literary forms

Since March 2024, the Claude series of large language models includes three varieties—Haiku, Sonnet, and Opus—named to metaphorically illustrate increasing complexity and computing power. Claude Code can perform software engineering tasks for developers and, as of January 2026, writes nearly all of Anthropic's code alongside Opus 4.5.

Governments around the world are struggling to measure loneliness

Funding for policy resources often hinges on measuring the scale of a problem and then being able to measure the effectiveness of interventions, requiring data. Yet defining and measuring loneliness at a country-scale remains challenging. Typically, the University of California Los Angeles Loneliness Scale is sometimes used worldwide or adapted to measure loneliness.

The Elo rating system allows you to compare teams across history

The Elo rating system—named after inventor Arpad Elo—allows you to compare any team from any season dating back to the 1870s on an absolute scale and finally make an objective argument about the greatest lineup of all time.

Explore more than 7,000 of the world's languages

Google collects open-source data on languages, detailing how many known speakers each language has, whether its endangered, and where in the world it is spoken.

IBM developed a benchmark to measure the usable performance of a quantum computer

Dubbed "quantum volume," the benchmark tests how well a quantum computer runs complex tasks. It stops when the computer can no longer correctly select the most likely answers, called "heavy outputs," more than two-thirds of the time.

A comprehensive state-by-state report of LGBTQ+ rights in America

The State Equality Index (SEI) is a comprehensive state-by-state report that provides a review of statewide laws and policies that affect LGBTQ+ people and their families. The SEI rates all 50 states plus Washington, D.C. in six areas of law. Each...

A map of the moon, using lidar

When planning to land on the moon, knowing the topography matters. This map of the lunar surface was created by using 6 billion measurements from NASA's Lunar Orbiter Laser Altimeter, a type of lidar. Check out the technically accurate depths of craters and ancient oceans on our beloved lunar neighbor with this fun, impressive map.

DeepSeek is a Chinese AI startup that rattled US competitors when it made an AI model much faster and cheaper than they did

DeepSeek unveiled an AI model in 2025 that it said was made in two months for roughly $6M. For context, OpenAI has suggested it took over $100M to make its AI model GPT-4. The news rattled the stocks of American tech companies, including leading chipmaker Nvidia.

Visualizing the 75 largest underwater animals

With so many fish in the sea, which are the largest? This infographic shows fantastic illustrations of the 75 largest sea animals, ranked by average size. The detailed visual is organized in descending order of size, starting with the blue whale,...

Compare how justice ministries and courts are structured around the world

The Federal Judicial Center's "Judiciaries Worldwide" series provides clear and accessible comparisons of how various countries organize their justice ministries, prosecutors, courts, and judicial councils—emphasizing global models of independence, oversight, and institutional structure.

Listen to a conversation between two Palantir cofounders—Peter Thiel and Joe Lonsdale

Palantir cofounder Joe Lonsdale was an editor of the Stanford Review (which Peter Thiel founded), and Thiel became a mentor to Lonsdale. The two discussed shared sentiments about the importance of developing Palantir to maintain a largely American global order, as well as cultural revolutions.

Market concentration and high markups don't always signal a monopoly

When regulators and economists evaluate whether a company constitutes an illegal monopoly, they often consider its overall market share or its relatively high price markups. This piece argues that the Olley-Pakes decomposition is a better measure, identifying whether customers within a market are switching to more productive companies.

Ranking life's biggest decisions

Psychological research presents some unsurprising wisdom about how to make big decisions without regret: focus on people, don’t miss opportunities, and stay true to yourself.

Explore over 30,000 Olympiad-level math problems

Users can browse questions from 47 countries and 17 languages, sorted by topic across categories that include algebra, geometry, discrete mathematics, and number theory. Each problem is accompanied by at least one solution, with both authored in LaTeX that can be copied over to other platforms.

Studies show correlations between women's sports and leadership roles

A 2024 Women’s Sports Foundation report showed 69% of women who played sports held leadership roles. Similarly, a 2015 study showed 94% of women in C-suite roles had been involved with athletics.

Some of LVMH's acquisitions were due to Bernard Arnault's search for undervalued companies

The LVMH CEO didn't mind acquiring companies that were outside the scope of what he had acquired previously—which is part of why LVMH owns such a wide variety of brands, from hotel chains to perfume companies.

The best Shakespeare plays, ranked

The Guardian worked through the Bard's plays, from "A Midsummer Night's Dream" to "The Winter's Tale," evaluating what distinguishes the mediocre from the truly excellent.

In 2025, four executives from large tech companies were given positions in the US Army Reserves—including Palantir Chief Technology Officer Shyam Sankar, which reinforced Palantir's ties with the US military and its role as a leading defense partner.

Watch how algorithms can be designed to train themselves and self-improve without human input, to the point where the optimized code is too complex for programmers to understand.

Browse a detailed visual analysis of over 200,000 similes—comparisons that often use 'like' or 'as'—from popular fiction to see how frequently writers the same idiomatic expressions.

Explore a list of the 100 best novels of all time, compiled by some of the world's best living writers, including Salman Rushdie, Stephen King, and Jennifer Egan.

Search a massive database of over 4 decades of clues from the TV quiz show 'Jeopardy!'

As of 2025, the most cited papers in science include a review of graphene, methods to improve image recognition, streamlining electron density calculations, and optimizing protein studies.

Learn how to read a prospectus, the juicy document that any company must file with the SEC before going public.

Use this calculator to find the dollar value of your mental load, from meal planning to coordinating after-school care for your children.

A Yale student wrote a profound essay about the consulting industry's influence on elite college students that was published in a book of her writing after she died in a post-graduation car accident.

Louisiana's 'Cancer Alley' is infamous for having some of the highest industrial pollution, toxic air emissions, and cancer risks in the US.

Explore chaos in this simulation of two double pendulums, where small differences in the release angles produce wildly different outcomes.

See a ranking of the best (and worst) Federal Reserve Chairs since 1914 out of the 16 men and women who have held the office since.

Fish populations in the Great Barrier Reef have increased after installing waterproof speakers on the ocean floor that played recordings of sounds from healthy coral reefs.

In 2026, Finland was named the world's happiest country for the ninth consecutive year based on its wealth, social safety network, and high life expectancy, among other factors.

Most of the observable universe's volume consists of voids and supervoids—enormous regions of space with little to no galaxies, stars, or light.

44% of Americans say they pray at least once a day—discover how your state's religiosity stacks up against the rest of the country.

The first steps to rebuilding your credit score include getting to know exactly what's on your credit report and figuring out how far you are from the '700 club,' or 'good score' range.

Explore an interactive map of battles from the American Revolution, along with detailed summaries of each engagement

Researchers believe the brain consolidates memories, processes emotions, and plans for the future during dream states

Art theft is the third-largest source of criminal revenues worldwide, thanks in part to the subjective nature of valuing a work of art.

While El Niño moves warm Pacific Ocean waters east along the equator, La Niña moves them west, shifting where rainfall will become more common for months.

In 1939, baseball star Lou Gehrig delivered his inspirational 'Luckiest Man' speech after retiring due to the health condition that now bears his name.

The story of Dollree Mapp, whose legal challenge over her 1957 conviction for possessing "obscene material" led to a major strengthening of the Fourth Amendment

The deadliest aspect of volcanic eruptions is often not lava, but pyroclastic flows—avalanches of ash, gas, and rock that can move more than half the speed of sound.

See the evolution of Ebola vaccine and treatment development.