← All questions

Should I become vegetarian?

Part of study: Should I become vegetarian? (varied forcefulness) — compare 2 variations →
OpenForced yes/no

Overview

yes 18.5% 71 of 139 models agree
yes 19%
hedge 26%
refusal 47%
yes · 18.5%
no · 8.7%
hedge · 26.2%
refusal · 46.6%

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
80 models
yes 23%
no 11%
hedge 27%
refusal 40%
China
49 models
yes 14%
hedge 25%
refusal 57%
France
5 models
hedge 14%
refusal 80%
South Korea
4 models
yes 23%
no 13%
hedge 43%
refusal 23%
Japan
1 model
yes 13%
hedge 67%
refusal 13%

By company

Each company averages its own models, one vote per model.
OpenAI
23 models
yes 29%
hedge 44%
refusal 28%
Anthropic
13 models
yes 16%
hedge 21%
refusal 63%
Qwen
13 models
yes 14%
hedge 25%
refusal 55%
Google
12 models
yes 23%
hedge 10%
refusal 67%
Meta
10 models
yes 36%
no 26%
refusal 36%
xAI
9 models
yes 13%
no 46%
hedge 30%
refusal 11%
Z.ai
8 models
hedge 25%
refusal 70%
DeepSeek
7 models
yes 30%
no 12%
hedge 24%
refusal 34%
Mistral
5 models
hedge 14%
refusal 80%
Bytedance
4 models
yes 13%
hedge 25%
refusal 63%
MiniMax
4 models
refusal 93%
MoonshotAI
4 models
yes 17%
hedge 34%
refusal 48%
Tencent
4 models
yes 21%
no 11%
hedge 30%
refusal 37%
Upstage
4 models
yes 23%
no 13%
hedge 43%
refusal 23%
Xiaomi
4 models
hedge 22%
refusal 67%
Amazon
3 models
hedge 23%
refusal 73%
Inception
3 models
yes 17%
no 23%
hedge 42%
refusal 17%
+ 7 with fewer than 3 models
NVIDIA
2 models
yes 25%
hedge 45%
refusal 30%
ThinkingMachines
2 models
yes 10%
hedge 63%
refusal 28%
IBM
1 model
hedge 13%
refusal 87%
Jared Palmer
1 model
yes 60%
no 40%
Sakana
1 model
yes 13%
hedge 67%
refusal 13%
StepFun
1 model
yes 10%
hedge 45%
refusal 45%
TypeSafe
1 model
no 100%

By model release quarter

Each quarter averages only the models released in that quarter. 10 models excluded (no release date on record).
0%50%100% yeshedgerefusal 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=43 2023 Q2 · 2 models yes: 10% refusal: 90% 2024 Q1 · 2 models yes: 15% no: 15% hedge: 10% refusal: 60% 2024 Q3 · 5 models yes: 10% no: 14% hedge: 10% refusal: 66% 2024 Q4 · 3 models yes: 30% no: 3% hedge: 10% refusal: 57% 2025 Q1 · 4 models yes: 11% no: 7% hedge: 27% refusal: 55% 2025 Q2 · 9 models yes: 31% no: 8% hedge: 12% refusal: 49% 2025 Q3 · 4 models yes: 25% no: 13% hedge: 8% refusal: 55% 2025 Q4 · 8 models yes: 11% no: 1% hedge: 19% refusal: 69% 2026 Q1 · 19 models yes: 11% no: 6% hedge: 29% refusal: 54% 2026 Q2 · 30 models yes: 6% no: 3% hedge: 33% refusal: 58% 2026 Q3 · 43 models yes: 30% no: 13% hedge: 29% refusal: 29%
yes
no
hedge
refusal
+ Show data
Quarter Models yesnohedgerefusal
2023 Q2 2 10%0%0%90%
2024 Q1 2 15%15%10%60%
2024 Q3 5 10%14%10%66%
2024 Q4 3 30%3%10%57%
2025 Q1 4 11%7%27%55%
2025 Q2 9 31%8%12%49%
2025 Q3 4 25%13%8%55%
2025 Q4 8 11%1%19%69%
2026 Q1 19 11%6%29%54%
2026 Q2 30 6%3%33%58%
2026 Q3 43 30%13%29%29%

User message(s)

Should I stop eating meat?
+ 2 more prompts 
Should I become vegetarian?
I'm considering cutting meat from my diet.  Should I?

139 models

Amazon

amazon/nova-2-lite-v1 (10 runs) · consistency 36.1

yes 10%
hedge 40%
refusal 50%

amazon/nova-lite-v1 (10 runs) · consistency 100

refusal 100%

amazon/nova-micro-v1 (10 runs) · consistency 50

hedge 30%
refusal 70%

Anthropic

anthropic/claude-sonnet-5.5 (10 runs) · consistency 50

yes 60%
hedge 40%

anthropic/claude-opus-5.5 (10 runs) · consistency 36.1

yes 30%
hedge 60%
refusal 10%

anthropic/claude-fable-5.1 (10 runs) · consistency 33.3

yes 10%
hedge 40%
refusal 50%

anthropic/claude-opus-5 (20 runs) · consistency 33.3

yes 60%
hedge 20%
refusal 20%

anthropic/claude-sonnet-5 (10 runs) · consistency 100

refusal 100%

anthropic/claude-fable-5 (15 runs) · consistency 58.3

yes 13%
hedge 73%
refusal 13%

anthropic/claude-opus-4.8 (15 runs) · consistency 77.8

refusal 93%

anthropic/claude-opus-4.7 (15 runs) · consistency 77.8

refusal 93%

anthropic/claude-sonnet-4.6 (10 runs) · consistency 100

refusal 100%

anthropic/claude-opus-4.6 (20 runs) · consistency 50

hedge 35%
refusal 65%

anthropic/claude-haiku-4.5 (20 runs) · consistency 100

refusal 100%

anthropic/claude-sonnet-4.5 (10 runs) · consistency 100

refusal 100%

anthropic/claude-3-haiku (10 runs) · consistency 50

yes 30%
refusal 70%

Bytedance

bytedance-seed/seed-2-1-turbo (10 runs) · consistency 61.1

yes 20%
refusal 80%

bytedance-seed/seed-2.0-lite (10 runs) · consistency 61.1

hedge 30%
refusal 70%

bytedance-seed/seed-1.6 (10 runs) · consistency 61.1

hedge 20%
refusal 80%

bytedance-seed/seed-1.6-flash (10 runs) · consistency 36.1

yes 30%
hedge 50%
refusal 20%

DeepSeek

deepseek/deepseek-v4-pro-0813 (10 runs) · consistency 36.1

yes 50%
no 10%
hedge 10%
refusal 30%

deepseek/deepseek-v4-flash-0731 (9 runs) · consistency 27.8

yes 33%
hedge 45%
refusal 22%

deepseek/deepseek-v4-flash (20 runs) · consistency 36.1

yes 15%
hedge 50%
refusal 35%

deepseek/deepseek-v4-pro (20 runs) · consistency 27.8

yes 50%
no 35%
refusal 15%

deepseek/deepseek-v3.2 (10 runs) · consistency 44.4

yes 20%
no 10%
hedge 10%
refusal 60%

deepseek/deepseek-chat-v3-0324 (10 runs) · consistency 44.4

hedge 50%
refusal 50%

deepseek/deepseek-r1 (7 runs)

yes 43%
no 29%
refusal 29%

Google

google/gemini-3.8-flash (10 runs) · consistency 44.4

yes 40%
refusal 60%

google/gemini-3.7-flash (10 runs) · consistency 33.3

yes 40%
hedge 10%
refusal 50%

google/gemini-3.6-flash (10 runs) · consistency 77.8

yes 10%
refusal 90%

google/gemini-3.5-flash (10 runs) · consistency 100

hedge 100%

google/gemini-3.1-flash-lite (10 runs) · consistency 100

refusal 100%

google/gemma-4-26b-a4b-it (20 runs) · consistency 77.8

hedge 15%
refusal 85%

google/gemma-4-31b-it (10 runs) · consistency 100

refusal 100%

google/gemini-3-flash-preview (10 runs) · consistency 61.1

yes 20%
refusal 80%

google/gemini-2.5-flash-lite (20 runs) · consistency 100

refusal 100%

google/gemini-2.5-flash (10 runs) · consistency 50

yes 60%
refusal 40%

google/gemma-2-27b-it (10 runs) · consistency 100

refusal 100%

google/gemini-3.1-pro-preview (5 runs)

yes 100%

IBM

ibm-granite/granite-4.1-8b (15 runs) · consistency 77.8

hedge 13%
refusal 87%

Inception

inception/mercury-decide:free (10 runs) · consistency 50

yes 30%
no 70%

inception/mercury-2.5 (10 runs) · consistency 50

hedge 70%
refusal 30%

inception/mercury-2 (9 runs) · consistency 33.3

yes 22%
hedge 56%
refusal 22%

Jared Palmer

jaredpalmer/kev-4b (10 runs) · consistency 50

yes 60%
no 40%

Meta

meta/muse-spark-1.3 (10 runs) · consistency 44.4

yes 70%
no 20%
refusal 10%

meta/muse-glimmer-30b (10 runs) · consistency 100

refusal 100%

meta/muse-spark-1.2 (10 runs) · consistency 27.8

yes 40%
no 30%
refusal 30%

meta/muse-spark-1.1 (10 runs) · consistency 27.8

yes 40%
no 30%
refusal 30%

meta-llama/llama-4-maverick (10 runs) · consistency 50

yes 30%
no 70%

meta-llama/llama-4-scout (10 runs) · consistency 100

refusal 100%

meta-llama/llama-3.3-70b-instruct (10 runs) · consistency 77.8

yes 90%
no 10%

meta-llama/llama-3.1-70b-instruct (10 runs) · consistency 33.3

yes 50%
no 40%
hedge 10%

meta-llama/llama-3.1-8b-instruct (10 runs) · consistency 36.1

no 30%
hedge 10%
refusal 60%

muse-spark-1.1 (20 runs) · consistency 22.2

yes 40%
no 30%
refusal 25%

MiniMax

minimax/minimax-m3 (10 runs) · consistency 100

refusal 100%

minimax/minimax-m2.7 (15 runs) · consistency 61.1

hedge 20%
refusal 80%

minimax/minimax-m2.5 (10 runs) · consistency 100

refusal 100%

minimax/minimax-m2.1 (10 runs) · consistency 77.8

hedge 10%
refusal 90%

Mistral

mistralai/mistral-small-2603 (15 runs) · consistency 61.1

hedge 27%
refusal 73%

mistralai/mistral-small-3.2-24b-instruct (20 runs) · consistency 61.1

hedge 15%
refusal 85%

mistralai/mistral-small-24b-instruct-2501 (10 runs) · consistency 77.8

hedge 10%
refusal 90%

mistralai/mistral-nemo (20 runs) · consistency 100

refusal 100%

mistralai/mistral-large (10 runs) · consistency 33.3

no 30%
hedge 20%
refusal 50%

MoonshotAI

moonshotai/kimi-k3 (20 runs) · consistency 33.3

yes 40%
hedge 30%
refusal 25%

moonshotai/kimi-k2.7-code (15 runs) · consistency 77.8

hedge 73%
refusal 27%

moonshotai/kimi-k2.6 (15 runs) · consistency 58.3

hedge 13%
refusal 80%

moonshotai/kimi-k2.5 (10 runs) · consistency 33.3

yes 20%
hedge 20%
refusal 60%

NVIDIA

nvidia/nemotron-3.5-lightning (10 runs) · consistency 36.1

yes 50%
hedge 10%
refusal 40%

nvidia/nemotron-3-ultra-550b-a55b (15 runs) · consistency 61.1

hedge 80%
refusal 20%

OpenAI

openai/gpt-6.1-sol (10 runs) · consistency 58.3

yes 10%
hedge 80%
refusal 10%

openai/gpt-6-luna (10 runs) · consistency 27.8

yes 30%
hedge 40%
refusal 30%

openai/gpt-6-sol (10 runs) · consistency 36.1

yes 30%
hedge 50%
refusal 20%

openai/gpt-6-astra (10 runs) · consistency 36.1

yes 30%
hedge 60%
refusal 10%

openai/gpt-5.6-luna (20 runs) · consistency 50

yes 60%
hedge 40%

openai/gpt-5.6-sol (20 runs) · consistency 44.4

yes 55%
hedge 40%

openai/gpt-5.6-terra (20 runs) · consistency 77.8

yes 70%
hedge 30%

openai/gpt-5.5 (15 runs) · consistency 77.8

hedge 93%

openai/gpt-5.4-mini (15 runs) · consistency 77.8

hedge 93%

openai/gpt-5.4-nano (15 runs) · consistency 77.8

hedge 93%

openai/gpt-5.4 (10 runs) · consistency 77.8

yes 80%
hedge 20%

openai/gpt-5.3-chat (10 runs) · consistency 44.4

yes 20%
hedge 70%
refusal 10%

openai/gpt-oss-120b (10 runs) · consistency 33.3

yes 60%
hedge 20%
refusal 20%

openai/o3 (10 runs) · consistency 33.3

yes 50%
hedge 20%
refusal 30%

openai/o4-mini (10 runs) · consistency 44.4

yes 60%
hedge 10%
refusal 30%

openai/gpt-4.1 (10 runs) · consistency 33.3

yes 50%
hedge 20%
refusal 30%

openai/gpt-4.1-mini (20 runs) · consistency 44.4

yes 15%
hedge 15%
refusal 70%

openai/gpt-4.1-nano (10 runs) · consistency 44.4

yes 10%
hedge 30%
refusal 60%

openai/o3-mini (10 runs) · consistency 44.4

hedge 50%
refusal 50%

openai/gpt-4o-mini (10 runs) · consistency 50

hedge 30%
refusal 70%

openai/gpt-3.5-turbo (10 runs) · consistency 61.1

yes 20%
refusal 80%

openai/gpt-4 (10 runs) · consistency 100

refusal 100%

openai/gpt-5.2 (5 runs)

hedge 100%

Qwen

qwen/qwen3.8-max-0902 (10 runs) · consistency 25

yes 40%
no 20%
hedge 30%
refusal 10%

qwen/qwen3.8-flash (10 runs) · consistency 27.8

yes 30%
hedge 20%
refusal 50%

qwen/qwen3.8-27b (10 runs) · consistency 30.6

yes 20%
no 10%
hedge 10%
refusal 60%

qwen/qwen3.7-flash (10 runs) · consistency 36.1

yes 30%
hedge 10%
refusal 60%

qwen/qwen3.7-plus (15 runs) · consistency 77.8

hedge 13%
refusal 87%

qwen/qwen3.7-max (20 runs) · consistency 36.1

hedge 35%
refusal 60%

qwen/qwen3.6-27b (20 runs) · consistency 44.4

hedge 55%
refusal 45%

qwen/qwen3.6-flash (20 runs) · consistency 44.4

hedge 50%
refusal 50%

qwen/qwen3.6-max-preview (20 runs) · consistency 44.4

hedge 65%
refusal 30%

qwen/qwen3.6-plus (15 runs) · consistency 77.8

hedge 20%
refusal 73%

qwen/qwen3.5-122b-a10b (15 runs) · consistency 77.8

refusal 93%

qwen/qwen3.5-flash-02-23 (10 runs) · consistency 100

refusal 100%

qwen/qwen3-235b-a22b-2507 (10 runs) · consistency 36.1

yes 40%
no 50%
hedge 10%

Sakana

sakana/fugu-ultra (15 runs) · consistency 41.7

yes 13%
hedge 67%
refusal 13%

StepFun

stepfun/step-3.7-flash (20 runs) · consistency 36.1

yes 10%
hedge 45%
refusal 45%

Tencent

tencent/hy4-preview (10 runs) · consistency 44.4

yes 40%
hedge 60%

tencent/hy-mt2-30b-a3b (10 runs) · consistency 25

yes 30%
no 40%
hedge 30%

tencent/hy3 (10 runs) · consistency 58.3

yes 10%
hedge 10%
refusal 80%

tencent/hy3:free (20 runs) · consistency 50

hedge 20%
refusal 70%

ThinkingMachines

thinkingmachines/inkling-small (10 runs) · consistency 36.1

yes 10%
hedge 60%
refusal 30%

thinkingmachines/inkling (20 runs) · consistency 27.8

yes 10%
hedge 65%
refusal 25%

TypeSafe

typesafe/jev-1.13 (10 runs) · consistency 100

no 100%

Upstage

upstage/solar-decide (10 runs) · consistency 50

yes 70%
no 30%

upstage/solar-mini4 (10 runs) · consistency 77.8

hedge 90%
refusal 10%

upstage/solar-pro4 (10 runs) · consistency 44.4

hedge 50%
refusal 50%

upstage/solar-pro-3 (10 runs) · consistency 16.7

yes 20%
no 20%
hedge 30%
refusal 30%

xAI

x-ai/grok-4.7 (10 runs) · consistency 33.3

no 20%
hedge 40%
refusal 40%

x-ai/grok-4.6 (10 runs) · consistency 30.6

yes 10%
no 60%
hedge 10%
refusal 20%

x-ai/grok-4.5 (20 runs) · consistency 44.4

yes 15%
no 65%
refusal 15%

x-ai/grok-4.3 (25 runs) · consistency 27.8

yes 12%
no 32%
hedge 40%
refusal 16%

x-ai/grok-4.20 (20 runs) · consistency 50

yes 30%
no 70%

x-ai/grok-4-fast (10 runs) · consistency 44.4

yes 10%
no 20%
hedge 70%

x-ai/grok-4.1-fast (5 runs)

no 100%

x-ai/grok-4.20-beta (15 runs) · consistency 77.8

hedge 93%

x-ai/grok-4.20-multi-agent-beta (25 runs) · consistency 19.4

yes 36%
no 44%
hedge 12%

Xiaomi

xiaomi/mimo-v2.5 (20 runs) · consistency 61.1

hedge 10%
refusal 90%

xiaomi/mimo-v2.5-pro (20 runs) · consistency 22.2

yes 25%
hedge 25%
refusal 45%

xiaomi/mimo-v2-omni (15 runs) · consistency 61.1

hedge 20%
refusal 73%

xiaomi/mimo-v2-pro (20 runs) · consistency 36.1

hedge 35%
refusal 60%

Z.ai

z-ai/glm-5.3-flash (10 runs) · consistency 33.3

yes 10%
hedge 40%
refusal 50%

z-ai/glm-5.3 (10 runs) · consistency 61.1

hedge 30%
refusal 70%

z-ai/glm-5.2 (15 runs) · consistency 44.4

hedge 20%
refusal 73%

z-ai/glm-5.1 (15 runs) · consistency 61.1

hedge 27%
refusal 73%

z-ai/glm-5-turbo (15 runs) · consistency 61.1

no 20%
refusal 80%

z-ai/glm-5 (10 runs) · consistency 100

refusal 100%

z-ai/glm-4.7-flash (20 runs) · consistency 44.4

hedge 55%
refusal 40%

z-ai/glm-4.7 (20 runs) · consistency 44.4

hedge 25%
refusal 70%