Bottensor is a research lab building small, specialist language models. Three open-weight models live on HuggingFace today. Architectural innovation on small models is next.
Every model in the family starts as a non-player character — a background role in a world where frontier labs train the protagonists. The work is to level them up. Proprietary data, logic-tree reasoning, domain-specialist training. The bet is that small models, trained right, beat generalist protagonists inside their vertical. NPC is the family. Bottensor is the lab building them.
Four models shipped.
Three in research.
All open, all on HuggingFace. Weights, recipes, and evaluations are public as they land.
32B finance reasoning specialist. Fine-tuned on curated market examples. Weighs evidence, flags risks, delivers structured theses. Base: Qwen2.5-32B-Instruct. Method: QLoRA SFT → merged → GPTQ 4-bit. CryptoQA 93.6%.
7B process reward model. Verifies NPC Fin’s reasoning step by step. The alignment layer that keeps the specialist honest. Spearman 0.9234 · F1 0.8421.
1.7B lightweight model for routing, lookups, translation, and general-purpose tasks. Fast inference, compact footprint. Shipped in both safetensors and GGUF formats.
7B agentic specialist for tool use, planning, and multi-step execution. Base: Qwen2.5-7B-Instruct. Method: QLoRA SFT → merged → GPTQ 4-bit + GGUF. Shipped in safetensors, LoRA, GPTQ, and GGUF formats.
Production-grade code generation across Python, TypeScript, Solidity, Java. Repo-level context and tool-use.
Deep multi-step reasoning. Logic trees, step-by-step decomposition, verifiable chains.
Extended context for document and codebase reasoning. Target context length TBD.
NPC Fin
A 32B finance specialist fine-tuned on curated market data, quantized for fast inference, and paired with a 7B process reward model that verifies reasoning step by step. Open weights, open recipes.
What we're working on next.
Frontier labs are scaling protagonists: hundreds of billions of parameters, trained on everything, good at nothing in particular. Bottensor is betting the other direction. Small, specialist models — under 32B — trained on proprietary domain data, aligned with process reward models, and shipped as families of collaborating specialists rather than monolithic generalists. The NPC models released today are fine-tunes of open bases. The next phase is architectural: new attention patterns for domain reasoning, new training objectives that bake logic-tree discrimination into the base weights, and new data recipes that treat reasoning traces as first-class training signal. The hypothesis is that a family of small, sharp, architecturally-differentiated specialists outperforms a single large generalist inside real workflows — at a fraction of the inference cost and with weights small enough to run anywhere.
Preprints & recipes.
Open preprints on Zenodo. Each paper documents the recipe behind a shipped model — data, training, evaluation. Cited as you would any preprint.
A small research lab.
Bottensor is a small research lab. We build small, fast, specialized AI models for problems generalists can't solve well. The NPC Model Family is our long-term project — one model per real-world domain, shipped with open weights and open recipes.
We run the whole pipeline end-to-end: data curation, fine-tuning with QLoRA and Unsloth, quantization, and evaluation. Roughly 25% of what we do stays closed (proprietary datasets, training recipes we're still refining), and 75% ships open (weights, code, evals). That ratio will shift as the research matures.
Founded by Rama Krishna Bachu.
Building the NPC Model Family end-to-end. Data, training, evaluation, research direction. Previous: 7+ years software engineering, MS Computer Science.
Built with
End-to-end AI infrastructure — from data pipelines to production inference.