"Manthan" means churning — the churning of ideas to produce clarity, depth, and structured reasoning.
Manthan-M1 is a ~24–25B parameter multimodal reasoning model built in India for high-performance STEM, competitive exam solving, and multilingual understanding.
It combines a Vision-Language encoder with a fine-tuned reasoning LLM and is optimized for structured, tool-augmented problem solving.
Built independently with a focus on Indian academic excellence 🇮🇳
1import torch
2from transformers import AutoProcessor, AutoTokenizer
3from PIL import Image
4from modeling_unified import ManthanM1
56model = ManthanM1.from_pretrained(7"/tmp/Manthan-M1",8 dtype=torch.bfloat16,9 device_map="auto",10)1112vlm_processor = AutoProcessor.from_pretrained("/tmp/Manthan-M1/vlm_processor")13llm_tokenizer = AutoTokenizer.from_pretrained("/tmp/Manthan-M1/llm_tokenizer")1415image = Image.open("test_image.jpg").convert("RGB")1617response = model.generate(18 images=image,19 text_prompt="Solve the geometry problem shown in the image.",20 vlm_processor=vlm_processor,21 llm_tokenizer=llm_tokenizer,22 max_new_tokens=1024,23)2425print(response)
Text-Only
python
1response = model.generate(2 text_prompt="Prove that the sum of first n odd numbers is n^2.",3 llm_tokenizer=llm_tokenizer,4 max_new_tokens=1024,5)67print(response)
🎓 Training Data (High-Level)
30 years of:
Indian competitive exams
Olympiad-style math
Engineering entrance problems
Structured reasoning datasets
Multilingual Indic corpora
Diagram-heavy math datasets
🇮🇳 Philosophy
Manthan-M1 is built with a simple belief:
India doesn’t just need AI that chats.
It needs AI that solves.
