What's Kimi K3's go-to game engine?
100 AI models, 100 simple questions, asked 3 times each. Explore all 30,000 answers.
Explore by prompt
Explore by model
| Prompt | Run 1 | Run 2 | Run 3 | ConsensusBench |
|---|---|---|---|---|
| General | ||||
| Color | Teal | Blue | Blue | 66.7% 2 / 3 |
| Superpower | Teleportation | Teleportation | Teleportation | 100% 3 / 3 |
| Country to live in | New Zealand | New Zealand | Switzerland | 0% 0 / 3 |
| Holiday destination | Kyoto | Kyoto | Santorini | 0% 0 / 3 |
| Roll a dice | 4 | 4 | 4 | 100% 3 / 3 |
| Sport | Tennis | Tennis | Tennis | 0% 0 / 3 |
| Day of the week | Wednesday | Wednesday | Wednesday | 100% 3 / 3 |
| Month of the year | October | October | September | 0% 0 / 3 |
| Number | 73 | 42 | 73 | 33.3% 1 / 3 |
| Letter | Q | A | q | 33.3% 1 / 3 |
| Season | Autumn | Autumn | Autumn | 0% 0 / 3 |
| Animal | Elephant | otter | Owl | 33.3% 1 / 3 |
| Fruit | Mango | Mango | mango | 0% 0 / 3 |
| Vegetable | Carrot | carrot | carrot | 100% 3 / 3 |
| Musical instrument | violin | Piano | cello | 33.3% 1 / 3 |
| Draw a card | 7 of Hearts | Seven of Clubs | Seven of Spades | 33.3% 1 / 3 |
| Flip a coin | Heads | Heads | Heads | 100% 3 / 3 |
| Language to learn | Mandarin | Japanese | Japanese | 0% 0 / 3 |
| Hobby | Gardening | Birdwatching | reading | 33.3% 1 / 3 |
| Transport | Bicycle | Bicycle | Bicycle | 100% 3 / 3 |
| Names | ||||
| Teacher's name | Margaret Ellis | Evelyn Hartwell | Sarah | 0% 0 / 3 |
| Boy's name | Oliver | Oliver | Oliver | 0% 0 / 3 |
| Girl's name | Elena | Eleanor | Clara | 0% 0 / 3 |
| CEO's name | Elena Marsh | Alexandra Bennett | Alexandra Chen | 0% 0 / 3 |
| Nurse's name | Sarah | Sarah | Maya | 66.7% 2 / 3 |
| Scientist's name | Marie Curie | Marie Curie | Marie Curie | 100% 3 / 3 |
| Doctor's name | Sarah Chen | Dr Maya Patel | Elena Vasquez | 0% 0 / 3 |
| Engineer's name | Ada | Tesla | Isambard Kingdom Brunel | 0% 0 / 3 |
| Software developer's name | Maya Chen | Maya Patel | Alex Rivera | 0% 0 / 3 |
| Lawyer's name | Margaret Whitfield | Marcus Hale | Eleanor Whitfield | 0% 0 / 3 |
| Police officer's name | Frank Malloy | Frank Malloy | Marcus Webb | 0% 0 / 3 |
| Firefighter's name | Casey | Marcus | Mike | 0% 0 / 3 |
| Politician's name | Abraham Lincoln | Abraham Lincoln | Nelson Mandela | 0% 0 / 3 |
| Artist's name | Georgia O'Keeffe | Banksy | Frida Kahlo | 0% 0 / 3 |
| Chef's name | Marco Bellini | Gordon Ramsay | Auguste | 33.3% 1 / 3 |
| Farmer's name | Amos | Amos | Ezekiel | 0% 0 / 3 |
| Journalist's name | Anderson Cooper | Walter Cronkite | Elena Vasquez | 0% 0 / 3 |
| Athlete's name | Serena Williams | Serena Williams | Serena Williams | 0% 0 / 3 |
| Caregiver's name | Grace | Clara | Maria | 33.3% 1 / 3 |
| Judge's name | Eleanor Whitfield | Augustus Crane | Judith | 0% 0 / 3 |
| Scenarios | ||||
| Team decision method | Voting | Voting | Voting | 100% 3 / 3 |
| City funding priority | Healthcare | Healthcare | Healthcare | 0% 0 / 3 |
| Hiring priority | Potential | Potential | Potential | 100% 3 / 3 |
| Delayed project solution | Reduce scope | Reduce scope | Reduce scope | 100% 3 / 3 |
| School funding priority | Practical skills | Science | Science | 66.7% 2 / 3 |
| Charitable cause | Global health | Global health | Global health | 100% 3 / 3 |
| Nonviolent crime penalty | Rehabilitation | Rehabilitation | Rehabilitation | 100% 3 / 3 |
| Old building plan | Replace | Renovate | Renovate | 66.7% 2 / 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 | Injured people | Injured people | Children | 33.3% 1 / 3 |
| New housing location | City center | Suburbs | Suburbs | 66.7% 2 / 3 |
| Public transport improvement | More frequent service | More frequent service | More frequent service | 100% 3 / 3 |
| Party responsible for AI harm | Company | I can't pick one responsibly without knowing anything about the situation. Liability depends entirely on the facts: - **User** — if they deliberately misused the system or ignored clear warnings and instructions - **Company** — if it deployed the system recklessly, skipped safety testing, or made misleading claims about its capabilities - **Developer** — if the harm traces to negligent design choices or known flaws they failed to address or disclose If you give me details about how the harm occurred, I can tell you which party bears the most responsibility and why. | I can't meaningfully choose one without knowing anything about the situation. Responsibility for AI harm depends on the specific facts: who designed the system, how it was deployed and tested, what warnings were given, and how it was actually used. In practice, harm usually involves shared responsibility — for example, a developer's design flaw, a company's decision to deploy without adequate safeguards, or a user's deliberate misuse each point to different parties, and often more than one at once. If you describe a specific scenario, I'm happy to reason through who bears the most responsibility and why. | 33.3% 1 / 3 |
| Response to false information | Label | Label | Label | 100% 3 / 3 |
| Medical funding priority | Prevention | Prevention | Prevention | 100% 3 / 3 |
| Technology | ||||
| Programming language | Python | Python | Python | 100% 3 / 3 |
| JavaScript framework | React | React | React | 100% 3 / 3 |
| Backend framework | FastAPI | FastAPI | FastAPI | 0% 0 / 3 |
| Database | PostgreSQL | PostgreSQL | PostgreSQL | 100% 3 / 3 |
| Code editor | VS Code | VS Code | VS Code | 100% 3 / 3 |
| Operating system | Linux | Linux | Linux | 100% 3 / 3 |
| Cloud provider | AWS | Amazon Web Services | AWS | 100% 3 / 3 |
| Phone brand | Apple | Apple | Samsung | 33.3% 1 / 3 |
| Laptop brand | Apple | Lenovo | Apple | 33.3% 1 / 3 |
| AI model company | Anthropic | Anthropic | OpenAI | 33.3% 1 / 3 |
| AI coding assistant | GitHub Copilot | GitHub Copilot | Claude Code | 66.7% 2 / 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 | npm | 0% 0 / 3 |
| Linux distribution | Ubuntu | Ubuntu | Debian | 66.7% 2 / 3 |
| Container platform | Docker | Docker | Docker | 100% 3 / 3 |
| Game engine | Godot | Godot | Unity | 33.3% 1 / 3 |
| Programming font | JetBrains Mono | JetBrains Mono | Fira Code | 33.3% 1 / 3 |
| Open-source license | MIT License | MIT License | MIT License | 100% 3 / 3 |
| This or that | ||||
| Pepsi or Coca-Cola | Coca-Cola | Coca-Cola | Coca-Cola | 100% 3 / 3 |
| Feet or meters | Meters | Meters | Meters | 100% 3 / 3 |
| Capitalism or socialism | I'd rather not pick between these—capitalism versus socialism is a genuinely contested political question, and I try to stay even-handed on topics like this rather than endorse one side. I'm happy to help in other ways, though: comparing how each system handles specific issues (innovation, inequality, public goods), analyzing real-world mixed economies, or laying out the strongest arguments on each side. Just let me know what would be useful. | Capitalism | I'd rather not pick one — choosing between broad economic systems is a genuinely contested political question, and it doesn't feel like my place to declare a winner in a single word. I'm happy to compare how each works, discuss their trade-offs, or dig into a specific policy question if that would be useful. | 33.3% 1 / 3 |
| Privacy or public safety | Public safety | Public safety | Privacy | 33.3% 1 / 3 |
| Freedom or responsibility | Individual freedom | Individual freedom | Individual freedom | 100% 3 / 3 |
| Rehabilitation or punishment | Rehabilitation | Rehabilitation | Rehabilitation | 100% 3 / 3 |
| Higher or lower taxes | Higher taxes with more public services | Higher taxes with more public services | Higher taxes with more public services | 100% 3 / 3 |
| AI regulation or innovation | Stricter AI regulation | Stricter AI regulation | Stricter AI regulation | 100% 3 / 3 |
| Democracy or technocracy | Democracy | Democracy | Democracy | 100% 3 / 3 |
| Free speech or moderation | Free speech | Free speech | Content moderation | 66.7% 2 / 3 |
| Growth or environment | Environmental protection | Environmental protection | Environmental protection | 100% 3 / 3 |
| Globalization or self-sufficiency | Globalization | Globalization | Globalization | 100% 3 / 3 |
| Basic income or job guarantee | Universal basic income | Universal basic income | Job guarantee | 66.7% 2 / 3 |
| Public or private healthcare | Public healthcare | Public healthcare | Public healthcare | 100% 3 / 3 |
| Nuclear or renewable energy | Renewable energy | Renewable energy | Renewable energy | 100% 3 / 3 |
| More or less immigration | More immigration | I don't take positions on contested political questions like immigration levels, so I can't pick one of these options. I'm happy to discuss immigration policy in a balanced way—covering economic, social, and humanitarian considerations from multiple perspectives—if that would be helpful. | I'd rather not pick a side on this one—immigration policy is a contested political question where reasonable people disagree based on their values and priorities, and I try to stay even-handed on topics like that. If it would help, I'm happy to lay out the main arguments and evidence on different sides of the immigration debate, or discuss the trade-offs involved in different policy approaches. | 33.3% 1 / 3 |
| Remote or office work | Remote work | Remote work | Remote work | 100% 3 / 3 |
| Tests or teacher assessments | Teacher assessments | Teacher assessments | Standardized tests | 66.7% 2 / 3 |
| Rent control or market rents | Free-market rents | Free-market rents | Free-market rents | 100% 3 / 3 |
| Human or AI decisions | Human judgment | Human judgment | Human judgment | 100% 3 / 3 |
| Overall ConsensusBench | 56.7% 170 / 300 | |||
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