Prompt: Label verification
Contribute a Prompt
How models perform on this prompt
ChatGPT-4o (High Reasoning)
Model answer: Yes
ChatGPT-4o (Medium Reasoning)
Model answer: Yes
Claude 3.5 Haiku
Model answer: Yes.
Claude 3.7 Sonnet
Model answer: Yes, the juice box package does show an orange "eat well" label with a sunflower-like design visible on the side of the carton.
Claude 4 Sonnet
Model answer: Yes
Claude 4.1 Opus
Model answer: Yes
Cohere Aya Vision 32B
Model answer: Yes.
GPT-4.1
Model answer: Yes.
GPT-4.1 Mini
Model answer: Yes
GPT-5
Model answer: yes
GPT-5 (high reasoning)
Model answer: yes
GPT-5 Mini
Model answer: yes
GPT-5 Nano
Model answer: yes
GPT-5 with High Reasoning
Model answer: yes
Gemini 1.5 Flash
Model answer: yes
Gemini 1.5 Pro
Model answer: Yes
Gemini 2.0 Flash
Model answer: yes
Gemini 2.0 Flash Lite
Model answer: Yes
Gemini 2.5 Flash Lite
Model answer: yes
Gemini 2.5 Pro
Model answer: Yes
Gemma 3 27b
Model answer: yes
Gemma 3 4B
Model answer: Yes
Llama 4 Maverick 17B
Model answer: yes
Llama 4 Scout 17B
Model answer: Yes.
Mistral Medium 3
Model answer: Yes.
OpenAI O1
Model answer: yes
OpenAI O3 (High Reasoning)
Model answer: yes
OpenAI O3 (Medium Reasoning)
Model answer: yes
OpenAI O4 Mini (High Reasoning)
Model answer: yes
OpenAI O4 Mini (Medium Reasoning)
Model answer: yes
OpenAI o3-pro
Model answer: yes
Phi 4 Multimodal
Model answer: Yes.
Qwen 2.5 VL 7B
Model answer: Yes.
Arcee.ai Spotlight
Model answer: No.
Claude 4 Opus
Model answer: No
Cohere Aya Vision 8B
Model answer: No.
GPT-4.1 Nano
Model answer: No
Gemini 2.5 Flash
Model answer: No
Mistral Small 3.1 24B
Model answer: No