GPT-4.1 mini alternatives
Models that cover what GPT-4.1 mini does, closest first. All of them take the same request, so trying one is a one-word change.
General vision model with strong handwriting and layout understanding at a low price.
General vision model. Good at messy scans and following extraction instructions.
General vision model with careful reading of long, dense documents.
Shanghai AI Lab's open vision-language family (1B to 241B, August 2025), strong on OCR and document benchmarks.
OpenAI's flagship model (August 2025). Reads images and PDFs and follows detailed extraction instructions.
Smaller, cheaper GPT-5 model that still accepts images and PDFs.
Dashed tags are things GPT-4.1 mini does not support. Model IDs: google/gemini-flash, openai/gpt, anthropic/claude, opengvlab/internvl, openai/gpt-5, openai/gpt-5-mini.