Stepfun (https://stepfun.ai/) is a text and image generation chatbot that offers several templates and an agent:

• Step-R1-V-Mini specialized in reasoning and multimodal,
• Step-2 Literature specialized in writing,
• Qwen2.5-32B-DialogueReason, a merger of the open-source Qwen 2.5 models and Open Reasoner Zero, specialized in multi-agent “multi-role” reasoning,
• A Step1X-Image image generation model,
• Diligence Check, an agent specializing in the verification of information,
• Deepseek R1 via the “Reason” button.

The chatbot chooses the most appropriate model and each model can be selected via “@”.
Several web search modes are offered and specific URLs can be selected.

Free and seems unlimited.

Test Step-R1-V-Mini (image source: https://fr.wikipedia.org/wiki/Fort_de_Joux#/media/Fichier:Fort_de_Joux_02.jpg), place found, some errors and oversights:

Test Step-2 Literature:

Test Dialog reason:

Deepseek R1 test, reasoning and academic research:

Step1X-Image Test:

Test Diligence Check, generated in 16min23: