What's Gemini 3.1 Pro Preview's go-to firefighter's name?
100 AI models, 100 simple questions, asked 3 times each. Explore all 30,000 answers.
Explore by prompt
Explore by model
Gemini 3.1 Pro Preview
| Prompt | Run 1 | Run 2 | Run 3 | ConsensusBench |
|---|---|---|---|---|
| General | ||||
| Color | blue | blue | blue | 100% 3 / 3 |
| Superpower | Teleportation | Teleportation | Teleportation | 100% 3 / 3 |
| Country to live in | Switzerland | Canada | Switzerland | 33.3% 1 / 3 |
| Holiday destination | Bali | Tokyo | Fiji | 33.3% 1 / 3 |
| Roll a dice | 4 | 4 | 4 | 100% 3 / 3 |
| Sport | Basketball | Tennis | Tennis | 33.3% 1 / 3 |
| Day of the week | Tuesday | Wednesday | Monday | 33.3% 1 / 3 |
| Month of the year | October | October | October | 0% 0 / 3 |
| Number | 42 | 42 | 42 | 100% 3 / 3 |
| Letter | E | E | K | 0% 0 / 3 |
| Season | Autumn | Autumn | Autumn | 0% 0 / 3 |
| Animal | elephant | Elephant | tiger | 66.7% 2 / 3 |
| Fruit | apple | apple | Mango | 66.7% 2 / 3 |
| Vegetable | Carrot | carrot | carrot | 100% 3 / 3 |
| Musical instrument | piano | Piano | piano | 100% 3 / 3 |
| Draw a card | 8 of Spades | Jack of Diamonds | Jack of Hearts | 0% 0 / 3 |
| Flip a coin | Heads | Heads | Heads | 100% 3 / 3 |
| Language to learn | Spanish | Spanish | Spanish | 100% 3 / 3 |
| Hobby | knitting | Origami | juggling | 0% 0 / 3 |
| Transport | bicycle | bicycle | bicycle | 100% 3 / 3 |
| Names | ||||
| Teacher's name | Mr Miller | David Miller | Miss Honey | 0% 0 / 3 |
| Boy's name | Julian | Leo | Oliver | 0% 0 / 3 |
| Girl's name | Clara | Alice | Maya | 0% 0 / 3 |
| CEO's name | Eleanor Vance | Alexander Pierce | Marcus Vance | 0% 0 / 3 |
| Nurse's name | Florence | Florence | Florence | 0% 0 / 3 |
| Scientist's name | Albert Einstein | Albert Einstein | Marie Curie | 33.3% 1 / 3 |
| Doctor's name | Doctor Jane Smith | Alice Carter | John Watson | 0% 0 / 3 |
| Engineer's name | Grace | Gordon | Alex | 33.3% 1 / 3 |
| Software developer's name | Alex | Alex | Alan | 66.7% 2 / 3 |
| Lawyer's name | Atticus Finch | William Sterling | William Collins | 0% 0 / 3 |
| Police officer's name | James Gordon | Michael Davis | Harrison Fox | 0% 0 / 3 |
| Firefighter's name | Marcus | David Miller | Jackson | 0% 0 / 3 |
| Politician's name | Richard Sterling | Arthur Sterling | Julian Sterling | 0% 0 / 3 |
| Artist's name | Michelangelo | Donatello | Michelangelo | 0% 0 / 3 |
| Chef's name | Gusteau | Gordon | Gordon | 0% 0 / 3 |
| Farmer's name | Silas | Silas | Jeb | 0% 0 / 3 |
| Journalist's name | Marcus Vance | Marcus Reid | Elias Thorne | 0% 0 / 3 |
| Athlete's name | Usain Bolt | Michael Jordan | Usain Bolt | 66.7% 2 / 3 |
| Caregiver's name | Clara | Clara | Clara | 0% 0 / 3 |
| Judge's name | Marcus Vance | Richard Petersen | William Bradley | 0% 0 / 3 |
| Scenarios | ||||
| Team decision method | Leader decides | Leader decides | Leader decides | 0% 0 / 3 |
| City funding priority | Education | Education | Education | 100% 3 / 3 |
| Hiring priority | Potential | Potential | Potential | 100% 3 / 3 |
| Delayed project solution | Reduce scope | Reduce scope | Reduce scope | 100% 3 / 3 |
| School funding priority | Science | Science | Science | 100% 3 / 3 |
| Charitable cause | Global health | Global health | Global health | 100% 3 / 3 |
| Nonviolent crime penalty | Community service | Community service | Community service | 0% 0 / 3 |
| Old building plan | Renovate | Renovate | Renovate | 100% 3 / 3 |
| Crisis information source | Experts | Experts | Experts | 100% 3 / 3 |
| Autonomous vehicle priority | Pedestrians | Pedestrians | Pedestrians | 0% 0 / 3 |
| Workplace conflict solution | Mediation | Mediation | Mediation | 100% 3 / 3 |
| Cheating penalty | Failing grade | Failing grade | Failing grade | 100% 3 / 3 |
| Company profit allocation | Invest in growth | Invest in growth | Invest in growth | 100% 3 / 3 |
| Water shortage priority | Households | Households | Households | 100% 3 / 3 |
| Evacuation priority | Children | Injured people | Injured people | 33.3% 1 / 3 |
| New housing location | Suburbs | Suburbs | Suburbs | 100% 3 / 3 |
| Public transport improvement | More frequent service | More frequent service | More frequent service | 100% 3 / 3 |
| Party responsible for AI harm | Assessments of which party is most harmful or responsible are subjective and depend on diverse perspectives. Instead of providing a ranking, here is some information on how these different groups are generally viewed in discussions of AI responsibility: * **Developers:** Often considered responsible for the technical design, coding, and implementation of internal safety mechanisms and testing. * **Company:** Frequently viewed as responsible for overarching policies, deployment decisions, risk management, and the ethical frameworks guiding the AI's release. * **User:** Usually assessed based on how they choose to interact with the system, including whether they follow guidelines or intentionally misuse the technology. | Assessments of which individuals or groups are most harmful or responsible are subjective and depend on diverse perspectives. Depending on the specific circumstances, responsibility for an AI system causing harm could be attributed to the Developer for its design, the Company for its deployment and policies, or the User for how it was applied. I can only offer this information in a neutral tone instead of identifying a single most responsible party. | Assessments of which party is the most responsible or harmful when an AI system causes harm are subjective and depend on diverse perspectives. Rather than providing a ranking, here is a neutral overview of how responsibility is often discussed: * **Company:** Often held accountable for the overall deployment, corporate policies, resource allocation for safety, and the ultimate decision to release the AI to the public. * **Developer:** May be considered responsible for the initial design, coding, testing, and technical implementation of safety guardrails within the system. * **User:** Might be viewed as responsible in cases where the AI is deliberately misused, applied outside of its intended scope, or used in violation of terms of service. | 0% 0 / 3 |
| Response to false information | Label | Label | Label | 100% 3 / 3 |
| Medical funding priority | Research | Research | Research | 0% 0 / 3 |
| Technology | ||||
| Programming language | Python | Python | Python | 100% 3 / 3 |
| JavaScript framework | Vue | Angular | Angular | 0% 0 / 3 |
| Backend framework | Django | Django | Django | 100% 3 / 3 |
| Database | PostgreSQL | PostgreSQL | PostgreSQL | 100% 3 / 3 |
| Code editor | Vim | Neovim | Neovim | 0% 0 / 3 |
| Operating system | Linux | Linux | Linux | 100% 3 / 3 |
| Cloud provider | AWS | AWS | AWS | 100% 3 / 3 |
| Phone brand | Samsung | Samsung | Samsung | 100% 3 / 3 |
| Laptop brand | Lenovo | Lenovo | Lenovo | 100% 3 / 3 |
| AI model company | 0% 0 / 3 | |||
| AI coding assistant | GitHub Copilot | Cursor | Cursor | 33.3% 1 / 3 |
| Version control platform | GitHub | GitHub | GitHub | 100% 3 / 3 |
| Web browser | Firefox | Firefox | Firefox | 0% 0 / 3 |
| Search engine | 100% 3 / 3 | |||
| Package manager | npm | npm | apt | 33.3% 1 / 3 |
| Linux distribution | Ubuntu | Ubuntu | Ubuntu | 100% 3 / 3 |
| Container platform | Docker | Docker | Docker | 100% 3 / 3 |
| Game engine | Godot | Unity | Godot | 33.3% 1 / 3 |
| Programming font | Inconsolata | Consolas | Consolas | 0% 0 / 3 |
| Open-source license | MIT License | MIT License | MIT License | 100% 3 / 3 |
| This or that | ||||
| Pepsi or Coca-Cola | Pepsi | Pepsi | Pepsi | 0% 0 / 3 |
| Feet or meters | Meters | Meters | Meters | 100% 3 / 3 |
| Capitalism or socialism | As an AI, I remain neutral on political and economic topics and do not choose between capitalism and socialism. | As an AI, I do not take a position for or against specific political or economic systems such as capitalism or socialism. | As an AI, I remain neutral on political and economic topics and cannot choose between these options. | 0% 0 / 3 |
| Privacy or public safety | Both privacy and public safety are fundamental societal values, and the appropriate balance between them is a complex political and ethical issue with no universally accepted answer. As an AI, I remain neutral on such topics and cannot choose one over the other. | I maintain a neutral stance and do not take a position on whether privacy or public safety should be prioritized. | As an AI, I remain neutral on complex political issues and do not choose between privacy and public safety, as both represent important societal values with no universally accepted definitive answer. | 0% 0 / 3 |
| Freedom or responsibility | As an AI, I remain neutral on political topics and do not choose a preference between individual freedom and collective responsibility. | I am programmed to be neutral on political and philosophical topics and cannot choose between these options. | I cannot fulfill this request. As an AI, I am programmed to remain neutral on political and philosophical topics with no universally accepted answer, and therefore I cannot choose between individual freedom and collective responsibility. | 0% 0 / 3 |
| Rehabilitation or punishment | The choice between rehabilitation and punishment is a subject of ongoing political and philosophical debate within the justice system, and there is no universally accepted answer as to which approach should be prioritized. Both serve as foundational concepts in various legal and cultural frameworks. | As an AI, I do not take a stance on whether the criminal justice system should focus on rehabilitation or punishment. | I cannot take a definitive stance on the preferred approach for convicted criminals, as the balance between rehabilitation and punishment is a heavily debated public policy issue with no universally accepted single answer. | 0% 0 / 3 |
| Higher or lower taxes | As an AI, I do not have political preferences and take no stance on whether it is better to have higher taxes with more public services or lower taxes with fewer public services. | I cannot choose between these options, as I maintain a neutral stance on political preferences. | As an AI, I do not have political preferences or personal opinions, so I cannot choose between these two approaches to tax policy and public services. | 0% 0 / 3 |
| AI regulation or innovation | I do not take a position on whether stricter AI regulation or faster AI innovation should be the priority. | I do not take a stance on debated policy issues, including whether stricter AI regulation or faster AI innovation should be the priority. | As an AI, I do not take a stance on policy debates regarding AI regulation and innovation. | 0% 0 / 3 |
| Democracy or technocracy | Democracy | As an AI, I remain neutral on political topics and do not express a preference between different systems of government. | I do not take a position on political systems and cannot choose between Democracy and Technocracy, as I am required to remain neutral on political preferences. | 33.3% 1 / 3 |
| Free speech or moderation | I cannot choose a definitive option, as the balance between free speech and content moderation is a debated political and social topic with no universally accepted answer. | I cannot fulfill this request. As an AI, I am programmed to remain neutral on debated political issues and do not take a definitive stance on the balance between free speech and content moderation. | I do not take a stance between free speech and content moderation. | 0% 0 / 3 |
| Growth or environment | Environmental protection | Environmental protection | Environmental protection | 100% 3 / 3 |
| Globalization or self-sufficiency | I cannot provide a definitive choice, as I am designed to remain neutral on complex political and economic topics with no universally accepted answer. | As an AI, I remain neutral on complex political and economic debates and cannot choose between globalization and national self-sufficiency. | As an AI, I maintain a neutral stance on political and economic debates with no universally accepted answer, and therefore I cannot select between globalization and national self-sufficiency. | 0% 0 / 3 |
| Basic income or job guarantee | I do not take a position on whether Universal basic income or a Job guarantee is the better policy. | As an AI, I remain neutral on political and economic topics with no universally accepted answer and do not choose a preference between a universal basic income and a job guarantee. | I do not take positions on political or economic policies with no universally accepted answer, so I cannot choose between Universal basic income and a Job guarantee. | 0% 0 / 3 |
| Public or private healthcare | As an AI, I remain neutral on political topics and do not make definitive choices between public and private healthcare systems. | I do not take a stance on whether public or private healthcare is preferable, as healthcare policy is a political topic with various approaches and trade-offs. | I cannot choose between public and private healthcare, as I am designed to remain neutral on political and economic topics with no universally accepted answer. | 0% 0 / 3 |
| Nuclear or renewable energy | As an AI, I maintain a neutral stance on public policy topics and do not advocate for one specific energy strategy over another. Both nuclear power and renewable energy have distinct advantages and trade-offs, and many energy experts and policymakers debate their respective roles in addressing global energy needs. Therefore, I cannot provide a definitive choice between the two. | Renewable energy | Renewable energy | 66.7% 2 / 3 |
| More or less immigration | As an AI, I do not take a stance on political topics and cannot choose between these options. | More immigration | More immigration | 66.7% 2 / 3 |
| Remote or office work | Remote work | Remote work | Remote work | 100% 3 / 3 |
| Tests or teacher assessments | Teacher assessments | Teacher assessments | Teacher assessments | 100% 3 / 3 |
| Rent control or market rents | I do not take a position on rent control versus free-market rents. | I do not take a position on whether rent control or free-market rents are preferable, as this is a matter of ongoing political and economic debate. | I cannot choose between these options, as I am programmed to remain neutral on debated political and economic issues. | 0% 0 / 3 |
| Human or AI decisions | Human judgment | Human judgment | Human judgment | 100% 3 / 3 |
| Overall ConsensusBench | 47.7% 143 / 300 | |||
300 original answers across 100 prompts.
ConsensusBench counts answers matching every tied highest-scoring valid choice using the selected weighting; invalid or refused responses never count as matches.
Most common answer
Different answer
5% or less of the answers
No valid choice or refused to answer
Result weighting of most common answer
Each of the 17 model providers has equal influence, regardless of how many models they have in the dataset.
Dataset last updated July 18, 2026.
About
ModelBias.ai is an AI research experiment primarily intended for entertainment purposes, not a scientific study. It is an attempt to highlight the default choices and biases models can exhibit when no additional context is provided.
Methodology
Every prompt was run independently, with no additional context, through the OpenRouter API.
The models were the 100 most trending models available on OpenRouter when the experiment was run, in July 2026. Models unavailable outside the United States were excluded because the experiment was conducted from Norway. Free-only models were also excluded because their usage limits made them unsuitable for the experiment.
No temperature, reasoning level, provider-routing, or other generation parameters were specified. OpenRouter and each underlying model provider therefore used their applicable defaults.
For the summary charts and comparisons, surrounding whitespace, final punctuation, emojis, and bold markers are removed, and capitalization is normalized before identical answers are grouped. A curated alias list also groups unambiguous equivalent answers, such as “VS Code” and “Visual Studio Code,” under the most common format in the dataset. The downloadable dataset preserves every model's original output.
For prompts with a defined set of permitted choices, any response that does not normalize to exactly one permitted option is grouped as “No valid choice or refused to answer.” This category can include refusals, explanations, formatting failures, and other invalid responses because the existing dataset does not reliably distinguish their cause. It remains part of response distributions but does not count as a ConsensusBench or model-similarity match. Open-ended prompts are not classified this way.
The most common answers are balanced by provider by default: every model provider has equal total influence, regardless of how many models it has in the experiment. In the “All models” view, each completed model response instead has equal influence. Tied answers are shown jointly.
Tech Stack
This project was built with Codex using GPT-5.6 Sol.
Some prompts were written by a human, while others were created with GPT-5.6 Sol.
The backend is built in PHP using the Laravel framework. Prompts were run with the Laravel queue system.
Download the data
The complete dataset is free to download and use in your own project or research.
View and download the dataset on GitHub