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.
Contrastive Language Models score a state against a set of candidate actions with a contrastive objective. Two projection heads on frozen Qwen3-8B embeddings, trained with InfoNCE; clm-serve maps Choice, Score and Noul onto candidate ranking.
Decides
choice, noul, score, classify, route
choice, score, noul, rank, classify, route
Architecture
nimble
clm
Fine-tuned from
qwen/qwen3.5-9b
qwen/qwen3-8b
License
apache-2.0
apache-2.0
Availability
Open weights + hosted API
Open weights
Hosted by
Bespoke Labs
—
Input price
—
—
Decision accuracy
90.1%
—
Calibration error
0.054
—
Valid action rate
—
—
Median latency
106 ms
—
p95 latency
—
—
Figures are from each model’s manifest; accuracy and latency are what the publishers report, on their own suites and hardware. Add a third model.
Questions
What is the difference between bespoke-nimble-9b and clm?
bespoke-nimble-9b is from Bespoke Labs and clm from Contrastive-LM. bespoke-nimble-9b has open weights and a hosted API; clm has open weights you can download and run. Both answer choice, noul, score, classify and route questions. Only clm answers rank. clm is the smaller model, at 8.0B parameters to 9.0B.
Which is more accurate, bespoke-nimble-9b or clm?
Only bespoke-nimble-9b publishes an accuracy figure (90.1% on Bespoke held-out set (324 examples)); clm does not, so there is no comparison to make without your own test.
Which is cheaper, bespoke-nimble-9b or clm?
bespoke-nimble-9b: Hosted, price not published, or free to self-host. clm: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Can I run bespoke-nimble-9b or clm locally?
Yes, both: systemone pull bespoke-labs/bespoke-nimble-9b and systemone pull contrastive-lm/clm download the weights.
