An agent makes hundreds of small decisions during a task. We built Mouse-1 to handle these decisions, saving you time and money. Its 163x cheaper, and 65x faster than GPT-6 Sol. It works by adding the model directly into your agent harness alongside your current stack. Mouse-1 takes the agent's state and options and returns a probability. This lets the harness handle all the small decisions while the main frontier model takes care of the reasoning or code.

Hey i'm Pete, I'm the guy that built Mouse. Adding a model like Mouse-1 to your agent harness will significantly improve performance, and cost savings. In our evals adding Mouse to an agent harness was made the agent ~190× cheaper and 13× faster compared to Opus 5.5 with the same exact output on testing. I optimized this model specifically for agent harnesses to perform fast decisions, classification, routing and more. Mouse-1 is more accurate and cheaper than Jev, and can be tuned and calibrated specifically for your agent harness, so it gets better the more you use it right from a single API endpoint. We pass through LoRa fine tuning from the h100 provider with zero mark up. The API is Zero Data Retention for all users by default. It's live and available to try today.

Here's a side by side of an agent harness with Mouse, and without vs GPT-6 Sol

set to the sweet sounds of Wu-Tang and Aesop Bach.

https://x.com/heypeterjames/status/2112919068705734748?s=20

Mouse-1

A decision model for your agent harness.

A decision model for your agent harness.

Forum Threads

what it do tho?

Forum Threads

what it do tho?