Optra Research
Small models, fine-tuned to do big work.
We fine-tune small language models and open-weight LLMs on your business, so you own your AI instead of renting it from someone else's cloud. We're not just distributing models: we also build apps on top of them, and every user those apps win becomes our distribution moat.
Our mission
AI for a billion Indians. Our mission is to put AI in the hands of a billion Indians.
Our vision
US frontier models are pricing out most of India: only a small fraction of Indian businesses and users can afford them. Open-weight models are capable enough for everyone else, but almost no one is offering them to everyday Indians. The models exist. What's missing is access.
How we do it differently
We have cracked distribution in India before: Chingari and the other Tech4Billion products reached hundreds of millions of Indians. Now we will do the same with Optra and Optra Research's products. We're bringing open-weight models into Optra's products, Optra Code, Optra Studio and more, to give hundreds of millions of Indians access to capable AI. We build the whole stack: our own GPUs, our own fine-tuned open-weight models and the apps on top. OptraCloud already runs a GPU farm of 50+ GPUs, owned by us and hosted on our own premises, serving all our apps today. We plan to grow it to 1,000+ GPUs in the next six months, and to thousands as demand scales. Because we own every layer, we can price for everyone: unlimited use, no per-token bills, no surprises.
Open by design
All our research is on open-weight models, AI that anyone can run, inspect and improve. That is how AI becomes truly local.
What we do
We fine-tune every model on our GPUs, 35 open-weight models across text, image, video, music, sound and voice, and we train small language models, for teams that need a model built for their own work.
These models are downloaded on the Optra GPUs. We are fine-tuning all of them.
For enterprises
Our products
What's next
Fine-tuning is where we start, not where we stop. As we grow, we will train our own models from the ground up: sovereign AI, built in India, for Indians. Few teams in India train their own models today. We intend to be one of the first to do it.
Contact
Talk to us: [email protected]. Tell us what your team needs a model to do.
If you are an ML researcher and want to join a cracked team which is making AI accessible for everyone, apply here to join.