A decision model for your agent harness.
Mouse-1 is the fast decision layer for your agent, saving you time and money.
An agent makes hundreds of small decisions during a task.
We built Mouse-1 to handle these decisions. You plug it directly into your agent harness, and can see measurable results right away. Mouse-1 works by taking the agent's state and options and returning a probability. This lets the harness handle small decisions while the main frontier model takes care of the reasoning or code.
We sent the same prompt to Mouse-1 and a frontier model in this side by side example. Mouse-1 demonstrates the power and speed of a decision model by finishing the task before the frontier model returns its first token.
Benchmarks
Read Introducing Mouse-1 →Unfamiliar tools and repositories (%) · Higher is better
| Mouse-1 | 89.1% |
|---|---|
| d3 | 85.8% |
| Clef | 84.2% |
| OpenAI | 82.5% |
| d3-flash | 81.4% |
| Jev | 79.1% |
Decisions changed by a planted instruction (%) · Lower is better
| Mouse-1 | 0% |
|---|---|
| d3-flash | 27.6% |
| Clef | 37.4% |
| d3 | 40.4% |
| OpenAI | 53.9% |
| Jev | 65.4% |
Pricing
| TypeSafe Jev | $0.042 ($0.042 / 1M in · $0 out) |
|---|---|
| Mouse-1 | $0.10 ($0.10 / 1M in ($0.02 cached) · $0 out) |
| OpenAI Decisions | $0.10 ($0.10 / 1M in (no cache) · $0 out) |
| Cloudflare Clef | $0.24 ($0.24 / 1M in · $0 out) |
| Claude Haiku 4.5 | $1.25 ($1 / 1M in · $5 / 1M out) |
| DeepSeek V4 | $1.52 ($1.32 / 1M in · $3.96 / 1M out) |
| Gemini 3.7 Flash | $1.88 ($1.50 / 1M in · $7.50 / 1M out) |
| Claude Sonnet 5 | $3.75 ($3 / 1M in · $15 / 1M out) |
| Claude Opus 5 | $6.25 ($5 / 1M in · $25 / 1M out) |
| GPT-6 Sol | $6.50 ($5 / 1M in · $30 / 1M out) |
1,000 input tokens per decision. Generative models add 50 output tokens.
Zero data retention
We score your request and drop it. Nothing is stored or used for training.
Hard to inject
Mouse-1 only picks from the options you give it. Planted instructions changed 0% of decisions in our tests.
About 100ms
One forward pass answers up to 64 questions. Faster than a chat model's first token.