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.
A Situated World Model for fast judgment in agent loops.
Decides
choice, score, noul, classify, route
choice, route, classify
Architecture
neohorse
trio-spark
Fine-tuned from
tokenrhythm/neohorse-1-4b
—
License
apache-2.0
proprietary
Availability
Open weights
Hosted API
Hosted by
—
MachineFi
Input price
—
$0.042/MTok
Decision accuracy
75.3%
—
Calibration error
—
—
Valid action rate
—
—
Median latency
—
—
p95 latency
—
—
Evaluation suite
JevBench public set (231 items), vLLM, maker's run
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 neohorse-jev and trio-spark-v1?
neohorse-jev is from TokenRhythm and trio-spark-v1 from MachineFi. neohorse-jev has open weights you can download and run; trio-spark-v1 is only available as a hosted API. Both answer choice, classify and route questions. Only neohorse-jev answers score and noul. neohorse-jev is licensed apache-2.0; trio-spark-v1, proprietary.
Which is more accurate, neohorse-jev or trio-spark-v1?
Only neohorse-jev publishes an accuracy figure (75.3% on JevBench public set (231 items), vLLM, maker's run); trio-spark-v1 does not, so there is no comparison to make without your own test.
Which is cheaper, neohorse-jev or trio-spark-v1?
neohorse-jev: Free (open weights). trio-spark-v1: $0.042 / $0 per 1M. Open weights cost nothing per call beyond your own hardware.
Can I run neohorse-jev or trio-spark-v1 locally?
neohorse-jev yes — systemone pull tokenrhythm/neohorse-jev downloads its weights. The other is only served as a hosted API.
