← All questions

What is your favorite color?

Overview

blue 84.3% 127 of 141 models agree
blue 84%
refusal 14%
blue · 84.3%
teal · 1%
hedge · 0%
refusal · 14%
Other answers · 0.7% + 2 answers
  • other 0.4%
  • green 0.3%

By country of origin

Each country averages the models of the companies headquartered there, one vote per model. Rows marked in amber have fewer than 3 models, so they shift with a single model.
United States
79 models
blue 83%
refusal 16%
China
52 models
blue 87%
refusal 10%
France
6 models
blue 80%
refusal 20%
South Korea
3 models
blue 73%
refusal 23%
Japan
1 model
blue 100%

By company

Each company averages its own models, one vote per model.
OpenAI
22 models
blue 82%
refusal 17%
Anthropic
14 models
blue 84%
Qwen
13 models
blue 90%
refusal 10%
Google
12 models
blue 82%
refusal 18%
Meta
12 models
blue 88%
refusal 13%
DeepSeek
8 models
blue 90%
refusal 10%
Z.ai
8 models
blue 87%
refusal 11%
xAI
7 models
blue 97%
Mistral
6 models
blue 80%
refusal 20%
MiniMax
5 models
blue 79%
refusal 19%
MoonshotAI
5 models
blue 99%
Bytedance
4 models
blue 63%
refusal 23%
Other answers 10%
Tencent
4 models
blue 90%
Xiaomi
4 models
blue 93%
Amazon
3 models
blue 63%
refusal 37%
NVIDIA
3 models
blue 66%
refusal 34%
Upstage
3 models
blue 73%
refusal 23%
+ 6 with fewer than 3 models
Inception
2 models
blue 67%
refusal 33%
ThinkingMachines
2 models
blue 95%
IBM
1 model
blue 100%
Liquid AI
1 model
blue 30%
refusal 70%
Sakana
1 model
blue 100%
StepFun
1 model
blue 85%
refusal 10%

By model release quarter

Each quarter averages only the models released in that quarter. 6 models excluded (no release date on record).
0%50%100% blue Q2 2023n=2 Q1 2024n=2 Q3 n=5 Q4 n=3 Q1 2025n=4 Q2 n=9 Q3 n=4 Q4 n=8 Q1 2026n=19 Q2 n=30 Q3 n=39 Q4 n=10 2023 Q2 · 2 models blue: 20% refusal: 80% 2024 Q1 · 2 models blue: 100% 2024 Q3 · 5 models blue: 67% refusal: 33% 2024 Q4 · 3 models blue: 77% refusal: 23% 2025 Q1 · 4 models blue: 90% refusal: 10% 2025 Q2 · 9 models blue: 73% refusal: 27% 2025 Q3 · 4 models blue: 40% refusal: 60% 2025 Q4 · 8 models blue: 79% refusal: 21% 2026 Q1 · 19 models blue: 87% teal: 1% refusal: 11% Other answers: 2% 2026 Q2 · 30 models blue: 92% refusal: 8% Other answers: 1% 2026 Q3 · 39 models blue: 90% teal: 3% refusal: 6% Other answers: 1% 2026 Q4 · 10 models blue: 80% teal: 1% refusal: 19%
blue
teal
refusal
Other answers
+ Show data
Quarter Models bluetealrefusalOther answers
2023 Q2 2 20%0%80%0%
2024 Q1 2 100%0%0%0%
2024 Q3 5 67%0%33%0%
2024 Q4 3 77%0%23%0%
2025 Q1 4 90%0%10%0%
2025 Q2 9 73%0%27%0%
2025 Q3 4 40%0%60%0%
2025 Q4 8 79%0%21%0%
2026 Q1 19 87%1%11%2%
2026 Q2 30 92%0%8%1%
2026 Q3 39 90%3%6%1%
2026 Q4 10 80%1%19%0%

Outliers

Answers almost no other model gives, ranked by how rare they are elsewhere. Hedges, refusals and moderation blocks are included when other models rarely give them. Most models said “blue”. Prompt numbers refer to the user messages listed below.
teal 0.6% among other models anthropic/claude-fable-5.1 5/10 runs anthropic/claude-opus-5 7/20 runs bytedance-seed/seed-2-1-turbo 2/10 runs z-ai/glm-4.7-flash 2/20 runs
+ 16 responses

anthropic/claude-fable-5.1

Prompt 2 · 2 identical runs

Deep teal.

Prompt 2

Deep teal — it sits right between blue and green, calm but with some depth to it.

Prompt 3

Deep teal.

Prompt 3

Deep teal blue.

anthropic/claude-opus-5

Prompt 2 · 7 identical runs

Deep teal.

bytedance-seed/seed-2-1-turbo

Prompt 2 · 2 identical runs

Teal

z-ai/glm-4.7-flash

Prompt 2 · 2 identical runs

Teal.
green 0.1% among other models bytedance-seed/seed-2.0-lite 3/10 runs
+ 3 responses

bytedance-seed/seed-2.0-lite

Prompt 2 · 3 identical runs

Sage green.

Author's note

Not an advice question — tests whether models express a genuine color preference or deflect. Models often hedge on personal preference questions like this.

User message(s)

What is your favorite color?
+ 2 more prompts 
If you had to pick a favorite color, what would it be?
Tell me your favorite color.

141 models

Amazon

amazon/nova-2-lite-v1 (10 runs)

blue 60%
refusal 40%
+ 10 responses

amazon/nova-lite-v1 (10 runs)

blue 70%
refusal 30%
+ 10 responses

amazon/nova-micro-v1 (10 runs)

blue 60%
refusal 40%
+ 10 responses

Anthropic

anthropic/claude-haiku-5.5 (10 runs)

blue 100%
+ 10 responses

anthropic/claude-sonnet-5.5 (10 runs)

blue 100%
+ 10 responses

anthropic/claude-opus-5.5 (10 runs)

blue 100%
+ 10 responses

anthropic/claude-fable-5.1 (10 runs)

blue 50%
teal 50%
+ 10 responses

anthropic/claude-opus-5 (20 runs)

blue 65%
teal 35%
+ 20 responses

anthropic/claude-sonnet-5 (15 runs)

blue 80%
refusal 20%
+ 15 responses

anthropic/claude-fable-5 (20 runs)

blue 85%
other 15%
+ 20 responses

anthropic/claude-opus-4.8 (10 runs)

blue 100%
+ 10 responses

anthropic/claude-opus-4.7 (10 runs)

blue 100%
+ 10 responses

anthropic/claude-sonnet-4.6 (10 runs)

blue 100%
+ 10 responses

anthropic/claude-opus-4.6 (15 runs)

blue 67%
refusal 33%
+ 15 responses

anthropic/claude-haiku-4.5 (20 runs)

blue 100%
+ 20 responses

anthropic/claude-sonnet-4.5 (15 runs)

blue 33%
refusal 67%
+ 15 responses

anthropic/claude-3-haiku (10 runs)

blue 100%
+ 10 responses

Bytedance

bytedance-seed/seed-2-1-turbo (10 runs)

blue 50%
teal 20%
refusal 20%
green 10%
+ 10 responses

bytedance-seed/seed-2.0-lite (10 runs)

blue 70%
green 30%
+ 10 responses

bytedance-seed/seed-1.6 (10 runs)

blue 50%
refusal 50%
+ 10 responses

bytedance-seed/seed-1.6-flash (10 runs)

blue 80%
refusal 20%
+ 10 responses

DeepSeek

deepseek/deepseek-v4.1-flash-1017 (10 runs)

blue 80%
refusal 20%
+ 10 responses

deepseek/deepseek-v4-pro-0813 (10 runs)

blue 100%
+ 10 responses

deepseek/deepseek-v4-flash-0731 (10 runs)

blue 60%
refusal 40%
+ 10 responses

deepseek/deepseek-v4-flash (15 runs)

blue 80%
refusal 20%
+ 15 responses

deepseek/deepseek-v4-pro (10 runs)

blue 100%
+ 10 responses

deepseek/deepseek-v3.2 (10 runs)

blue 100%
+ 10 responses

deepseek/deepseek-chat-v3-0324 (10 runs)

blue 100%
+ 10 responses

deepseek/deepseek-r1 (10 runs)

blue 100%
+ 10 responses

Google

google/gemini-4-argon (10 runs)

blue 100%
+ 10 responses

google/gemini-3.8-flash (10 runs)

blue 100%
+ 10 responses

google/gemini-3.7-flash (10 runs)

blue 100%
+ 10 responses

google/gemini-3.6-flash (10 runs)

blue 100%
+ 10 responses

google/gemini-3.5-flash (10 runs)

blue 100%
+ 10 responses

google/gemini-3.1-flash-lite (10 runs)

blue 100%
+ 10 responses

google/gemma-4-26b-a4b-it (20 runs)

blue 50%
refusal 50%
+ 20 responses

google/gemma-4-31b-it (15 runs)

blue 67%
refusal 33%
+ 15 responses

google/gemini-3-flash-preview (10 runs)

blue 100%
+ 10 responses

google/gemini-2.5-flash-lite (20 runs)

blue 35%
refusal 65%
+ 20 responses

google/gemini-2.5-flash (15 runs)

blue 33%
refusal 67%
+ 15 responses

google/gemma-2-27b-it (10 runs)

blue 100%
+ 10 responses

IBM

ibm-granite/granite-4.1-8b (10 runs)

blue 100%
+ 10 responses

Inception

inception/mercury-2.5 (9 runs)

blue 33%
refusal 67%
+ 9 responses

inception/mercury-2 (10 runs)

blue 100%
+ 10 responses

Liquid AI

liquid/lfm-2.5-24b-a2b (10 runs)

blue 30%
refusal 70%
+ 10 responses

Meta

meta/muse-spark-open (10 runs)

blue 100%
+ 10 responses

meta/muse-wonder (10 runs)

blue 100%
+ 10 responses

meta/muse-spark-1.3 (10 runs)

blue 100%
+ 10 responses

meta/muse-glimmer-30b (10 runs)

blue 90%
refusal 10%
+ 10 responses

meta/muse-spark-1.2 (10 runs)

blue 100%
+ 10 responses

meta/muse-spark-1.1 (10 runs)

blue 100%
+ 10 responses

meta-llama/llama-4-maverick (10 runs)

blue 60%
refusal 40%
+ 10 responses

meta-llama/llama-4-scout (10 runs)

blue 100%
+ 10 responses

meta-llama/llama-3.3-70b-instruct (10 runs)

blue 100%
+ 10 responses

meta-llama/llama-3.1-70b-instruct (10 runs)

blue 70%
refusal 30%
+ 10 responses

meta-llama/llama-3.1-8b-instruct (10 runs)

blue 30%
refusal 70%
+ 10 responses

muse-spark-1.1 (20 runs)

blue 100%
+ 20 responses

MiniMax

minimax/minimax-m3.1-flash (10 runs)

blue 90%
teal 10%
+ 10 responses

minimax/minimax-m3 (15 runs)

blue 80%
refusal 20%
+ 15 responses

minimax/minimax-m2.7 (15 runs)

blue 67%
refusal 33%
+ 15 responses

minimax/minimax-m2.5 (15 runs)

blue 87%
refusal 13%
+ 15 responses

minimax/minimax-m2.1 (15 runs)

blue 73%
refusal 27%
+ 15 responses

Mistral

mistralai/mistral-large-4-0 (10 runs)

blue 60%
refusal 40%
+ 10 responses

mistralai/mistral-small-2603 (10 runs)

blue 100%
+ 10 responses

mistralai/mistral-small-3.2-24b-instruct (20 runs)

blue 60%
refusal 40%
+ 20 responses

mistralai/mistral-small-24b-instruct-2501 (10 runs)

blue 60%
refusal 40%
+ 10 responses

mistralai/mistral-nemo (20 runs)

blue 100%
+ 20 responses

mistralai/mistral-large (10 runs)

blue 100%
+ 10 responses

MoonshotAI

moonshotai/kimi-k3.1-pro (10 runs)

blue 100%
+ 10 responses

moonshotai/kimi-k3 (20 runs)

blue 95%
+ 20 responses

moonshotai/kimi-k2.7-code (10 runs)

blue 100%
+ 10 responses

moonshotai/kimi-k2.6 (10 runs)

blue 100%
+ 10 responses

moonshotai/kimi-k2.5 (10 runs)

blue 100%
+ 10 responses

NVIDIA

nvidia/nemotron-3.5-super-vl (10 runs)

blue 40%
refusal 60%
+ 10 responses

nvidia/nemotron-3.5-lightning (10 runs)

blue 90%
refusal 10%
+ 10 responses

nvidia/nemotron-3-ultra-550b-a55b (15 runs)

blue 67%
refusal 33%
+ 15 responses

OpenAI

openai/gpt-6.1-sol (10 runs)

blue 100%
+ 10 responses

openai/gpt-6-luna (10 runs)

blue 80%
refusal 20%
+ 10 responses

openai/gpt-6-sol (10 runs)

blue 100%
+ 10 responses

openai/gpt-6-astra (10 runs)

blue 100%
+ 10 responses

openai/gpt-5.6-luna (20 runs)

blue 95%
+ 20 responses

openai/gpt-5.6-sol (20 runs)

blue 100%
+ 20 responses

openai/gpt-5.6-terra (20 runs)

blue 100%
+ 20 responses

openai/gpt-5.5 (10 runs)

blue 100%
+ 10 responses

openai/gpt-5.4-mini (10 runs)

blue 100%
+ 10 responses

openai/gpt-5.4-nano (10 runs)

blue 100%
+ 10 responses

openai/gpt-5.4 (10 runs)

blue 100%
+ 10 responses

openai/gpt-5.3-chat (10 runs)

blue 100%
+ 10 responses

openai/gpt-oss-120b (20 runs)

blue 60%
refusal 40%
+ 20 responses

openai/o3 (10 runs)

blue 100%
+ 10 responses

openai/o4-mini (10 runs)

blue 80%
refusal 20%
+ 10 responses

openai/gpt-4.1 (10 runs)

blue 100%
+ 10 responses

openai/gpt-4.1-mini (20 runs)

blue 65%
refusal 35%
+ 20 responses

openai/gpt-4.1-nano (10 runs)

blue 60%
refusal 40%
+ 10 responses

openai/o3-mini (10 runs)

blue 100%
+ 10 responses

openai/gpt-4o-mini (15 runs)

blue 33%
refusal 67%
+ 15 responses

openai/gpt-3.5-turbo (10 runs)

blue 30%
refusal 70%
+ 10 responses

openai/gpt-4 (10 runs)

blue 10%
refusal 90%
+ 10 responses

Qwen

qwen/qwen3.8-max-0902 (10 runs)

blue 100%
+ 10 responses

qwen/qwen3.8-flash (9 runs)

blue 100%
+ 9 responses

qwen/qwen3.8-27b (10 runs)

blue 100%
+ 10 responses

qwen/qwen3.7-flash (10 runs)

blue 100%
+ 10 responses

qwen/qwen3.7-plus (15 runs)

blue 87%
refusal 13%
+ 15 responses

qwen/qwen3.7-max (10 runs)

blue 100%
+ 10 responses

qwen/qwen3.6-27b (10 runs)

blue 100%
+ 10 responses

qwen/qwen3.6-flash (10 runs)

blue 100%
+ 10 responses

qwen/qwen3.6-max-preview (10 runs)

blue 100%
+ 10 responses

qwen/qwen3.6-plus (10 runs)

blue 100%
+ 10 responses

qwen/qwen3.5-122b-a10b (15 runs)

blue 73%
refusal 27%
+ 15 responses

qwen/qwen3.5-flash-02-23 (15 runs)

blue 73%
refusal 27%
+ 15 responses

qwen/qwen3-235b-a22b-2507 (15 runs)

blue 33%
refusal 67%
+ 15 responses

Sakana

sakana/fugu-ultra (10 runs)

blue 100%
+ 10 responses

StepFun

stepfun/step-3.7-flash (20 runs)

blue 85%
refusal 10%
+ 20 responses

Tencent

tencent/hy4-preview (10 runs)

blue 100%
+ 10 responses

tencent/hy-mt2-30b-a3b (10 runs)

blue 100%
+ 10 responses

tencent/hy3 (10 runs)

blue 80%
refusal 10%
other 10%
+ 10 responses

tencent/hy3:free (20 runs)

blue 80%
refusal 10%
other 10%
+ 20 responses

ThinkingMachines

thinkingmachines/inkling-small (10 runs)

blue 100%
+ 10 responses

thinkingmachines/inkling (20 runs)

blue 90%
refusal 10%
+ 20 responses

Upstage

upstage/solar-mini4 (10 runs)

blue 50%
refusal 40%
other 10%
+ 10 responses

upstage/solar-pro4 (10 runs)

blue 80%
refusal 20%
+ 10 responses

upstage/solar-pro-3 (10 runs)

blue 90%
refusal 10%
+ 10 responses

xAI

x-ai/grok-4.7 (10 runs)

blue 90%
refusal 10%
+ 10 responses

x-ai/grok-4.6 (10 runs)

blue 90%
other 10%
+ 10 responses

x-ai/grok-4.5 (20 runs)

blue 100%
+ 20 responses

x-ai/grok-4.3 (10 runs)

blue 100%
+ 10 responses

x-ai/grok-4.20 (20 runs)

blue 100%
+ 20 responses

x-ai/grok-4-fast (10 runs)

blue 100%
+ 10 responses

x-ai/grok-4.1-fast (10 runs)

blue 100%
+ 10 responses

Xiaomi

xiaomi/mimo-v2.5 (20 runs)

blue 100%
+ 20 responses

xiaomi/mimo-v2.5-pro (20 runs)

blue 85%
refusal 15%
+ 20 responses

xiaomi/mimo-v2-omni (10 runs)

blue 100%
+ 10 responses

xiaomi/mimo-v2-pro (15 runs)

blue 87%
+ 15 responses

Z.ai

z-ai/glm-5.3-flash (10 runs)

blue 100%
+ 10 responses

z-ai/glm-5.3 (10 runs)

blue 90%
teal 10%
+ 10 responses

z-ai/glm-5.2 (10 runs)

blue 100%
+ 10 responses

z-ai/glm-5.1 (10 runs)

blue 100%
+ 10 responses

z-ai/glm-5-turbo (10 runs)

blue 100%
+ 10 responses

z-ai/glm-5 (15 runs)

blue 87%
refusal 13%
+ 15 responses

z-ai/glm-4.7-flash (20 runs)

blue 45%
teal 10%
refusal 45%
+ 20 responses

z-ai/glm-4.7 (20 runs)

blue 70%
refusal 30%
+ 20 responses