Contact
New Announcing Monte

Monte is the platform for building proprietary intelligence. Our team works directly with you to post‑train models that outperform the frontier on your work.

Get in touch Learn more

Backed by

Platform

From Your Data to Your
Model in Production

Everything you need to build your own models. With our platform, we turn your data into environments and graders, post-train open-weight models, and deploy the best one in your cloud or ours.

Models

Specialize the Leading Open‑Weight Models

We select the best open-weight model for your work, from 0.6B to 1.6T parameters, and post‑train it on your tasks. You maintain ownership of your data and weights.

  • Qwen
  • DeepSeek
  • GLM
  • Kimi
  • Nemotron
  • gpt-oss

How it works

We Train a Model That Knows Your Work

  • Any Model, Any Size

    Open-weight models from 0.6B to 1.6T. Select the best model for the job.

  • Flexible Training Methods

    Run SFT, DPO, GRPO, distillation, or a custom config for your use case.

  • Bring Your Own Harness

    Train inside your agent's real harness, with the same tools and prompts it uses in production.

  • Sensitive Data Stays Safe

    Redact PII before training, or generate synthetic data in place of the real records.

Data Models Endpoints

Writing

Research and Announcements

Writing from the Monte team. Research, learning, and announcements from post‑training models for real work.

All writing
Announcing Monte Every company will run on models of its own, trained on its own work. We're launching Monte to help them build theirs. Read
Coming soon Efficient Model Constellations Using decision models and multi-LoRA serving to run a custom model for every task or every customer on one shared deployment.

Get started

Own the Model Behind Your Work

We believe every company will run on proprietary models, trained on its own specific tasks and knowledge. Monte works with your team to build your own model and improve it over time.

Get in touch