Small open typed-decision model with calibrated probabilities, first-class Nepali coverage.
TokenRhythm's prefill-only decision model for agent workflows. Built on its NeoHorse-1-4B (a Qwen3.5-4B derivative), it predicts Choice, Noul and Score answers over application-defined options without generating text, with optional single-image input.
Decides
choice, score, noul, classify, route
choice, score, noul, classify, route
Architecture
mira
neohorse
Fine-tuned from
jhu-clsp/mmbert-small
tokenrhythm/neohorse-1-4b
License
apache-2.0
apache-2.0
Availability
Open weights
Open weights
Hosted by
—
—
Input price
—
—
Decision accuracy
74.5%
75.3%
Calibration error
0.025
—
Valid action rate
100.0%
—
Median latency
30 ms
—
p95 latency
43 ms
—
Evaluation suite
s1-decision-bench
JevBench public set (231 items), vLLM, maker's run
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 mira and neohorse-jev?
mira is from SAGEA and neohorse-jev from TokenRhythm. Both have open weights you can download and run. Both answer choice, score, noul, classify and route questions.
Which is more accurate, mira or neohorse-jev?
They report on different suites — mira 74.5% on s1-decision-bench, neohorse-jev 75.3% on JevBench public set (231 items), vLLM, maker's run — so the numbers do not rank them. Test both on your own labelled examples.
Which is cheaper, mira or neohorse-jev?
mira: Free (open weights). neohorse-jev: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Can I run mira or neohorse-jev locally?
Yes, both: systemone pull sagea/mira and systemone pull tokenrhythm/neohorse-jev download the weights.
