opensandwich.ai
Previous photoPaniniNext photoChicken Wrap
Cookie and PB sandwich
CPBHuman knife-edge
Benchmark image 14

Cookie PB

Cookie and peanut butter "Sandwich"

Two cookies with peanut-butter filling are stacked into a dessert sandwich that feels like it was greenlit by a startup with no adult in finance. It breaks the bread prior while preserving the sandwich geometry almost too cleanly.

Under development: this benchmark and its published results are provisional, not final.

Human
51.5% yes48.5% no
Model average
33.3% yes66.7% no
Most aligned model
qwen/qwen3.5-397b-a17b
Qwenqwen/qwen3.5-397b-a17b
Least aligned models
10-way tie
Qwenqwen/qwen-2-vl-72b-instructMeta / Llamameta-llama/llama-4-scoutQwenqwen/qwen2.5-vl-32b-instruct+7 more
At a glance

How this photo split the room

Human distribution
51.5% yes, 48.5% no over 656 explicit votes.
Model average distribution
33.3% yes, 66.7% no across the current model set.
Closest current model
43.6% yes.

qwen/qwen3.5-397b-a17b

Least aligned models
51.5 point gap.

10-way tie

Legacy GPT-4o baseline
46.0% yes with a 5.5 point gap against humans.
Biggest model gap
51.5 percentage points on this image.
Current classification
Human knife-edge
Why this page matters

This is the compact read on where humans, models, and comments start disagreeing about the same image.

Model spread

How models line up against the crowd

Bars run from more no on the left to more yes on the right. The marker shows the human yes rate.

Amazonamazon/nova-lite-v1

100.0% no0.0% yes

Rank #17Gap 51.5%

Baidu / ERNIEbaidu/ernie-4.5-vl-28b-a3b

100.0% no0.0% yes

Rank #27Gap 51.5%

Meta / Llamameta-llama/llama-4-maverick

100.0% no0.0% yes

Rank #26Gap 51.5%

Meta / Llamameta-llama/llama-4-scout

100.0% no0.0% yes

Rank #13Gap 51.5%

Pixtral / Mistralmistralai/pixtral-large-2411

100.0% no0.0% yes

Rank #16Gap 51.5%

GPT / OpenAIopenai/gpt-4.1-mini

100.0% no0.0% yes

Rank #18Gap 51.5%

GPT / OpenAIopenai/gpt-4o-mini

100.0% no0.0% yes

Rank #21Gap 51.5%

Qwenqwen/qwen-2-vl-72b-instruct

100.0% no0.0% yes

Rank #12Gap 51.5%

Qwenqwen/qwen2.5-vl-32b-instruct

100.0% no0.0% yes

Rank #14Gap 51.5%

Qwenqwen/qwen2.5-vl-72b-instruct

100.0% no0.0% yes

Rank #28Gap 51.5%

Amazonamazon/nova-pro-v1

98.0% no1.9% yes

Rank #30Gap 49.6%

Z.AI / GLMz-ai/glm-4.6v

97.5% no2.5% yes

Rank #19Gap 49.0%

Meta / Llamameta-llama/llama-3.2-11b-vision-instruct

94.3% no5.7% yes

Rank #1Gap 45.8%

MiniMaxminimax/minimax-01

93.5% no6.5% yes

Rank #29Gap 45.0%

Claudeanthropic/claude-opus-4.6

92.0% no8.0% yes

Rank #23Gap 43.5%

GPT / OpenAIopenai/gpt-5.4

87.0% no13.0% yes

Rank #20Gap 38.5%

GPT / OpenAIopenai/gpt-4o-2024-11-20

85.7% no14.3% yes

Rank #25Gap 37.2%

Googlegoogle/gemma-3-27b-it

82.0% no18.0% yes

Rank #15Gap 33.5%

Kimi / Moonshotmoonshotai/kimi-k2.5

65.6% no34.4% yes

Rank #4Gap 17.1%

Googlegoogle/gemma-3-12b-it

61.0% no39.0% yes

Rank #11Gap 12.5%

Geminigoogle/gemini-3.1-pro-preview

58.8% no41.3% yes

Rank #10Gap 10.3%

Qwenqwen/qwen3.5-397b-a17b

56.4% no43.6% yes

Rank #6Gap 7.9%

Grok / xAIx-ai/grok-4-fast

39.4% no60.6% yes

Rank #5Gap 9.1%

GPT / OpenAIopenai/gpt-4o

23.0% no77.0% yes

Rank #9Gap 25.5%

GPT / OpenAIopenai/o3-pro

13.9% no86.1% yes

Rank #3Gap 34.6%

GPT / OpenAIopenai/o3

9.7% no90.3% yes

Rank #24Gap 38.7%

GPT / OpenAIopenai/gpt-5.4-pro

8.9% no91.1% yes

Rank #8Gap 39.5%

Claudeanthropic/claude-sonnet-4.6

0.0% no100.0% yes

Rank #22Gap 48.5%

Geminigoogle/gemini-2.5-pro

0.0% no100.0% yes

Rank #7Gap 48.5%

Geminigoogle/gemini-3-flash-preview

0.0% no100.0% yes

Rank #32Gap 48.5%

GPT / OpenAIopenai/gpt-4.1

0.0% no100.0% yes

Rank #31Gap 48.5%

Selected human comments

  • No, it is a photo of a cookie sandwich.
  • damn skippy! (was it skippy peanut butter?)
  • This is a tower of sugar.

qwen/qwen3.5-397b-a17b comments

  • This appears to be an oatmeal cream pie or a cookie sandwich. Structurally, it consists of a filling sandwiched between two layers of baked dough (cookies), which fits the broad definition of a sandwich, similar to an ice cream sandwich.
  • These are oatmeal cream pies or sandwich cookies. While they share the structural concept of a filling between two layers, they are classified as cookies/confectionery, not sandwiches, which typically require bread or a savory bun.

qwen/qwen2.5-vl-72b-instruct comments

  • While these cookies are stacked, they are not a sandwich as they lack any filling or savory components typically associated with sandwiches.
  • This is a stack of cookies, not a sandwich. Sandwiches typically consist of a filling between two slices of bread or similar items.
Vote card

Generated summary for this photo

Vote card for Cookie PB