# Bespoke Labs: bespoke-nimble-9b

> An open Jev-style LoRA on Qwen3.5-9B from Bespoke Labs, trained on 2,676 contrastively curated examples to score the allowed answer tokens directly for enums, booleans and rubric levels. Recipe, data and a public benchmark suite are released with it.

- Page: https://systemonemodels.ai/bespoke-labs/bespoke-nimble-9b
- API: https://api.systemonemodels.ai/v1/models/bespoke-labs/bespoke-nimble-9b
- Download: `pip install systemonemodels && systemone pull bespoke-labs/bespoke-nimble-9b`

## Facts

| | |
|---|---|
| Maker | Bespoke Labs (https://systemonemodels.ai/bespoke-labs) |
| Decides | choice, noul, score, classify, route |
| Architecture | nimble |
| Base model | qwen/qwen3.5-9b |
| Parameters | 9.0B |
| Context | 8K tokens |
| Licence | apache-2.0 |
| Availability | Open weights + hosted API |
| Released | 2026-09-18 |
| Latest version | 2026.09 |

## Hosted API

- Provider: Bespoke Labs (https://bespokelabs.ai)
- Docs: https://github.com/bespokelabsai/nimble
- Input: — · Output: —

## Reported evaluation

Suite: Bespoke held-out set (324 examples). Numbers are the publisher's own.

- Decision accuracy: 90.1%
- Calibration error (ECE): 0.054
- Median latency: 106 ms

## Model card

<!-- generated by scripts/seed_catalog.py; edit content/models/catalog.yaml -->

# Bespoke-Nimble-9B

An open Jev-style LoRA on Qwen3.5-9B from Bespoke Labs, trained on 2,676 contrastively curated examples to score the allowed answer tokens directly for enums, booleans and rubric levels. Recipe, data and a public benchmark suite are released with it.

On Bespoke's own 324-example held-out set Nimble scores 90.1% (Jev 93.2%) with an expected calibration error of 0.054 after the 22 September temperature fit; across 13 human-labelled public subsets it reaches 74.8% macro (Jev 76.0%). The 24 September checkpoint serves 8,192 tokens and 255 choices and ships with temperature 1.0; the earlier checkpoint is revision original-2676. The repository holds the adapter (about 165 MiB), not the base weights. A public demo without a key runs on Modal and speaks /v1/systemone; median latency there is 106 ms on an H100.

## What it decides

- **choice** — picks one option from a set
- **noul** — answers a yes/no question with one calibrated probability
- **score** — places the input on an ordered scale
- **classify** — assigns a category from a fixed taxonomy
- **route** — sends the input to one of several destinations

## At a glance

| | |
|---|---|
| Parameters | 9B |
| Base model | `qwen/qwen3.5-9b` |
| Maker | Bespoke Labs |
| Released | 2026-09-18 |
| License | apache-2.0 |
| Reported accuracy | 90.1% |
| Reported latency | 106 ms median on H100 |

## Hosted API

Served by **Bespoke Labs**. [Get access](https://bespokelabs.ai) · [API docs](https://github.com/bespokelabsai/nimble).

## Get the weights

```bash
pip install systemonemodels
systemone pull bespoke-labs/bespoke-nimble-9b
```

The files are served from the maker's Hugging Face repository, [`bespokelabs/Bespoke-Nimble-9B`](https://huggingface.co/bespokelabs/Bespoke-Nimble-9B), and verified against the checksums recorded here.

## Read more

- [Source, data and recipe](https://github.com/bespokelabsai/nimble)
- [Weights on Hugging Face](https://huggingface.co/bespokelabs/Bespoke-Nimble-9B)

---

*This page was opened by System One for Bespoke Labs, who can claim the organisation and take it over at any time.*

---

From System One Models — https://systemonemodels.ai/ · every System One model: https://systemonemodels.ai/system-one-models