Do we have OCR or Vision Models in Developer account? Is there any plan for it like DeepSeek OCR, IBM Granite Docling, Paddle OCR?
Please let me know if there is any plan?
Thank you for reaching out to us. SambaNova offers the Llama-4-Maverick-17B-128E-Instruct model, which supports vision and OCR capabilities, including document parsing, image captioning, and text extraction from scanned files.
You can find more details and usage examples in our documentation here: sambanova_doc
3 Likes
@ritesh.khapre there is no immediate plan for the OCR specific models at this time. However, we can file an enhancement request to get in the PM Teams line of sight .
@neha.marathe1 can you submit one idea, in product feedback, for each of the OCR models mentioned
-Coby
2 Likes
@coby.adams i see we have huge list of models Hume-Llama- but all are text. We need vision Model badly any basic to get started will be helpful and going forwarding ( hoping soon) OCR models as well and lately Time series models. We want Sambanova to be inference for everything.
On other note, will be helpful if we have package solution pricing listed in website to bring our model in your infra and how that process works, instead calling Sales Contact and the initiative goes to sink.