This new tool lets you train AI models in minutes
Thinking Machines Lab, started by former OpenAI CTO Mira Murati, just launched Tinker—a tool that lets developers fine-tune big language models without the usual tech headaches.
Tinker runs Python training loops on their distributed GPUs and handles all the behind-the-scenes stuff like scheduling and resource management, so users can focus on building cool AI projects.
Tinker is currently in private beta
Tinker is in private beta right now—researchers on the waitlist can use it for free, but a pay-as-you-go system is coming soon.
It works with several major AI models (like Alibaba's Qwen-235B-A22B), uses LoRA to keep computing costs down, and even includes an open-source Cookbook for post-training help.
Murati says she wants Tinker to encourage open research in AI.
The startup has already raised $2 billion in seed funding, showing investors are pretty excited about its future.