Bespoke Nimble
Hosted and open decision models that return a probability for every answer
Nimble is Bespoke Labs' decision model. Send the state and the answers you allow (choices or true/false), and it returns the answer and a probability for every option. It is built on Qwen3.8-27B for the current hosted version, nimble-v4-max.
The hosted API charges $0.04 per 1M input tokens from prepaid credit, output free, and new accounts get $1 of credit. Open weights and the training data and recipe are on Hugging Face and GitHub; the 27B weights are licensed CC BY-NC 4.0, which bars commercial use.
Pricing: Per token usage
Choosing a decision model right now?
One click, anonymous. Helps us know our visitors.
Thanks, that helps.
Choosing a decision model right now?
One click, anonymous. Helps us know our visitors.
Thanks, that helps.
Bespoke Nimble prices by model
Per 1M tokens, read off Bespoke Nimble's own pricing page on the date shown.
| Model | Input / 1M | Cached / 1M | Output / 1M | Checked | Notes |
|---|---|---|---|---|---|
| Nimble-v4-max | $0.04 | – | $0.00 | 25 Oct 2026 | Prepaid credit; new accounts get $1. Output tokens are free. |
Bespoke Nimble Alternatives
Explore 152 products in the Inference APIs category. View all Bespoke Nimble alternatives.
Kev
Open decision models from 0.8B to 27B, built on Qwen and compatible with TypeSafe's Jev API.
Laya
Open-source multilingual decision model: typed choices, scores and yes/no answers in one forward pass.
Mistral
Use models in a few clicks with our platform. Download our open models for deep access.
Work on Bespoke Nimble? Feature it at the top of Inference APIs.
Is your product missing?