Announcing Tinker

TinkerToy Computer invented by Daniel Hillis and Brian Silverman

Today, we are launching Tinker, a flexible API for fine-tuning language models. It empowers researchers and hackers to experiment with models by giving them control over the algorithms and data while we handle the complexity of distributed training. Tinker advances our mission of enabling more people to do research on cutting-edge models and customize them to their needs.

Tinker lets you fine-tune a range of large and small open-weight models, including large mixture-of-experts models such as Qwen-235B-A22B. Switching from a small model to a large one is as simple as changing a single string in your Python code.

Tinker is a managed service that runs on our internal clusters and training infrastructure. We handle scheduling, resource allocation, and failure recovery. This allows you to get small or large runs started immediately, without worrying about managing infrastructure. We use LoRA so that we can share the same pool of compute between multiple training runs, lowering costs.

Tinker’s API gives you low-level primitives like forward_backward and sample, which can be used to express most common post-training methods. Even so, achieving good results requires getting many details right. That’s why we’re releasing an open-source library, the Tinker Cookbook, with modern implementations of post-training methods that run on top of the Tinker API.

Groups at Princeton, Stanford, Berkeley, and Redwood Research have already been using Tinker:

• The Princeton Goedel Team trained mathematical theorem provers
• The Rotskoff Chemistry group at Stanford fine-tuned a model to complete chemistry reasoning tasks
• Berkeley’s SkyRL group ran experiments on a custom async off-policy RL training loop with multi-agents and multi-turn tool-use.
• Redwood Research used Tinker to RL Qwen3-32B on difficult AI control tasks

Tinker is now in private beta for researchers and developers. You can sign up for the Tinker waitlist here. We will be onboarding users to the platform starting today.

If you’re an organization interested in using Tinker, please contact us here.

Tinker will be free to start. We will introduce usage-based pricing in the coming weeks.

We’re excited to see what you discover and make with Tinker!