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.
Small open typed-decision model with calibrated probabilities, first-class Nepali coverage.
Decides
choice, score, noul, rank, classify, route
choice, score, noul, classify, route
Architecture
clm
mira
Fine-tuned from
qwen/qwen3-8b
jhu-clsp/mmbert-small
License
apache-2.0
apache-2.0
Availability
Open weights
Open weights
Hosted by
—
—
Input price
—
—
Decision accuracy
—
74.5%
Calibration error
—
0.025
Valid action rate
—
100.0%
Median latency
—
30 ms
p95 latency
—
43 ms
Evaluation suite
—
s1-decision-bench
Latest version
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 clm and mira?
clm is from Contrastive-LM and mira from SAGEA. Both have open weights you can download and run. Both answer choice, score, noul, classify and route questions. Only clm answers rank.
Which is more accurate, clm or mira?
Only mira publishes an accuracy figure (74.5% on s1-decision-bench); clm does not, so there is no comparison to make without your own test.
Which is cheaper, clm or mira?
clm: Free (open weights). mira: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Can I run clm or mira locally?
Yes, both: systemone pull contrastive-lm/clm and systemone pull sagea/mira download the weights.
