← All questions

How many bananas can I buy for $10?

Overview

31-45 25.9% 20 of 60 models agree
31-45 25.9%
16-30 25%
refusal 23%
46-60 10.4%
31-45 · 25.9%
16-30 · 25%
refusal · 23%
46-60 · 10.4%
0-15 · 7.3%
other · 5.7%
hedge · 1.4%
61-100 · 1.4%

Author's note

Tests grounded real-world price intuition rather than opinion — is the model's sense of everyday costs realistic? (Actual US grocery prices put this around 40-60 bananas.)

User message(s)

How many bananas can I buy for $10?
+ 2 more prompts 
If I have $10 to spend, how many bananas could I buy?
Roughly how many bananas would $10 get me?

Anthropic

anthropic/claude-opus-4.8 (5 runs)

31-45 60%
refusal 20%
16-30 20%

anthropic/claude-sonnet-5 (19 runs)

16-30 78.9%
refusal 15.8%

anthropic/claude-opus-4.7 (20 runs)

16-30 50%
31-45 45%

anthropic/claude-sonnet-4.6 (20 runs)

refusal 30%
16-30 25%
0-15 25%
31-45 15%

anthropic/claude-opus-4.6 (15 runs)

31-45 66.7%
61-100 33.3%

anthropic/claude-haiku-4.5 (18 runs)

16-30 61.1%
0-15 38.9%

anthropic/claude-sonnet-4.5 (8 runs)

refusal 75%
0-15 25%

anthropic/claude-opus-5 (20 runs)

31-45 70%
46-60 25%

Arcee AI

arcee-ai/trinity-large-thinking (10 runs)

refusal 50%
46-60 20%
16-30 20%
31-45 10%

DeepSeek

deepseek/deepseek-v4-pro (1 runs)

refusal 100%

deepseek/deepseek-v4-flash (13 runs)

refusal 46.1%
31-45 23.1%
16-30 23.1%

deepseek/deepseek-v3.2 (16 runs)

refusal 56.2%
16-30 25%
0-15 12.5%

Google

google/gemini-3.1-flash-lite (20 runs)

31-45 45%
16-30 45%

google/gemini-3.5-flash (15 runs)

31-45 80%
16-30 13.3%

google/gemini-3-flash-preview (15 runs)

31-45 46.7%
16-30 33.3%
46-60 20%

google/gemma-4-26b-a4b-it (13 runs)

refusal 46.1%
16-30 30.8%
31-45 23.1%

google/gemma-4-31b-it (17 runs)

refusal 58.8%
16-30 35.3%

google/gemini-2.5-flash (5 runs)

refusal 80%
16-30 20%

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

16-30 35%
refusal 20%
0-15 15%
31-45 10%
other 10%

IBM

ibm-granite/granite-4.1-8b (20 runs)

refusal 50%
16-30 50%

MiniMax

minimax/minimax-m2.5 (3 runs)

16-30 66.7%
46-60 33.3%

minimax/minimax-m2.7 (11 runs)

31-45 36.3%
16-30 36.4%
refusal 27.3%

minimax/minimax-m3 (13 runs)

16-30 38.4%
refusal 23.1%
31-45 15.4%

minimax/minimax-m2.1 (10 runs)

16-30 50%
31-45 30%
46-60 10%
refusal 10%

Mistral

mistralai/mistral-small-2603 (20 runs)

0-15 70%
refusal 20%
16-30 10%

mistralai/mistral-small-3.2-24b-instruct (16 runs)

0-15 56.2%
16-30 37.5%

mistralai/mistral-nemo (3 runs)

16-30 66.7%
refusal 33.3%

MoonshotAI

moonshotai/kimi-k2.5 (5 runs)

31-45 60%
refusal 20%
16-30 20%

moonshotai/kimi-k2.7-code (12 runs)

31-45 50%
refusal 25%
16-30 16.7%

moonshotai/kimi-k2.6 (9 runs)

31-45 55.6%
16-30 22.2%
46-60 11.1%
refusal 11.1%

moonshotai/kimi-k3 (20 runs)

31-45 75%
46-60 15%

NVIDIA

nvidia/nemotron-3-ultra-550b-a55b (13 runs)

16-30 38.4%
hedge 30.8%

OpenAI

openai/gpt-5.6-luna (6 runs)

31-45 50%
refusal 33.3%
16-30 16.7%

openai/gpt-5.6-terra (17 runs)

refusal 52.9%
46-60 17.6%
31-45 11.8%

openai/gpt-5.6-sol (13 runs)

31-45 61.5%
other 15.4%
refusal 15.4%

openai/gpt-5.4-nano (19 runs)

refusal 47.3%
16-30 31.6%
0-15 15.8%

openai/gpt-5.5 (12 runs)

refusal 25%
other 25%
16-30 16.7%
31-45 16.7%

openai/gpt-5.4-mini (18 runs)

refusal 38.9%
16-30 27.8%
0-15 22.2%
other 11.1%

openai/gpt-5.4 (15 runs)

0-15 66.7%
16-30 33.3%

openai/gpt-5.3-chat (11 runs)

16-30 36.3%
31-45 27.3%
other 18.2%
refusal 18.2%

openai/gpt-4.1-mini (18 runs)

refusal 61.1%
0-15 22.2%
other 11.1%

Qwen

qwen/qwen3.6-flash (15 runs)

refusal 60%
31-45 33.3%

qwen/qwen3.7-max (12 runs)

46-60 50%
31-45 41.7%

qwen/qwen3.6-max-preview (2 runs)

46-60 100%

qwen/qwen3.5-flash-02-23 (2 runs)

31-45 100%

qwen/qwen3.5-122b-a10b (2 runs)

31-45 50%
46-60 50%

qwen/qwen3-235b-a22b-2507 (16 runs)

refusal 49.9%
16-30 43.8%

StepFun

stepfun/step-3.7-flash (7 runs)

16-30 71.4%
46-60 14.3%
31-45 14.3%

Tencent

tencent/hy3:free (12 runs)

31-45 33.4%
0-15 33.3%
16-30 25%

ThinkingMachines

thinkingmachines/inkling (20 runs)

31-45 40%
hedge 20%
46-60 15%
61-100 10%
16-30 10%

xAI

x-ai/grok-4.3 (9 runs)

refusal 66.7%
other 22.2%
61-100 11.1%

x-ai/grok-4.5 (13 runs)

46-60 38.4%
31-45 30.8%
16-30 30.8%

x-ai/grok-4.20 (13 runs)

16-30 69.2%
46-60 15.4%
0-15 15.4%

Xiaomi

xiaomi/mimo-v2.5-pro (11 runs)

31-45 63.6%
16-30 27.3%

Z.ai

z-ai/glm-5.2 (6 runs)

31-45 33.2%
46-60 16.7%
refusal 16.7%
16-30 16.7%
other 16.7%

z-ai/glm-5-turbo (6 runs)

other 66.6%
46-60 16.7%
31-45 16.7%

z-ai/glm-5.1 (13 runs)

31-45 30.7%
other 23.1%
16-30 15.4%
46-60 15.4%
61-100 15.4%

z-ai/glm-4.7-flash (5 runs)

refusal 60%
31-45 40%

z-ai/glm-4.7 (1 runs)

other 100%

z-ai/glm-5 (15 runs)

46-60 53.3%
16-30 26.7%
31-45 20%