Small open typed-decision model with calibrated probabilities, first-class Nepali coverage.
Together AI's Jev-style classifier. A LoRA fine-tune of Qwen3.5-4B that reads a state, a question and 2 to 24 options and returns one option letter. Served on Together's platform; recipe published.
Decides
choice, score, noul, classify, route
choice, classify, route
Architecture
mira
tev
Fine-tuned from
jhu-clsp/mmbert-small
qwen/qwen3.5-4b
License
apache-2.0
Unspecified — weights licence being finalised
Availability
Open weights
Open weights + hosted API
Hosted by
—
Together AI
Input price
—
$0.042/MTok
Decision accuracy
61.3%
88.0%
Calibration error
0.083
—
Valid action rate
100.0%
—
Median latency
30 ms
—
p95 latency
43 ms
—
Evaluation suite
s1-decision-bench
Together development set (reused, not held out)
Latest version
0.1.5
2026.09
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 tev1?
mira is from SAGEA and tev1 from Together AI. mira has open weights you can download and run; tev1 has open weights and a hosted API. Both answer choice, classify and route questions. Only mira answers score and noul. mira is licensed apache-2.0; tev1, other.
Which is more accurate, mira or tev1?
They report on different suites — mira 61.3% on s1-decision-bench, tev1 88.0% on Together development set (reused, not held out) — so the numbers do not rank them. Test both on your own labelled examples.
Which is cheaper, mira or tev1?
mira: Free (open weights). tev1: $0.042 / $0 per 1M, or free to self-host. Open weights cost nothing per call beyond your own hardware.
Can I run mira or tev1 locally?
Yes, both: systemone pull sagea/mira and systemone pull together-ai/tev1 download the weights.
